(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Transactions of the Royal Society of Edinburgh"

TRANSACTIONS 



OF THE 



EOYAL SOCIETY OF EDINBUKGH. 



fit..tfM 



TRANSACTION S 



OP THE 



ROYAL SOCIETY 



OP 



EDINBURGH. 



VOL. XXX. 




EDINBURGH: 

PUBLISHED BY ROBERT GRANT & SON, 107 PRINCES STREET, 
AND WILLIAMS & NORGATE, 14 HENEIETTA STREET, COVEN T GARDEN, LONDON. 



MDCCCLXXXIII. 



CONTENTS. 



PART I. (1880-81.) 



PAGE 



I. The Law of Extensible Minors in Determinants. By Thomas 

Mum, M.A., ....... 1 

II. On some Transformations connecting General Determinants with 

Continuants. By Thomas Mum, M.A., ... 5 

III. Report on Fossil Fishes collected by the Geological Survey of Scot- 

land in Eskdale and Liddesdale. Part I. — Ganoidei. By 
Eamsay H. Traquair, M.D., F.R.S., Keeper of the Natural 
History Collections in the Museum of Science and Art, Edin- 
burgh. (Plates I. to VI., which will be given in the next 
Part), ........ 15 

IV. On some New Crustaceans from the Lower Carboniferous Rocks of 

Eskdale and Liddesdale. By B. N. Peach, A.R.S.M., F.G.S., 
of the Geological Survey of Scotland. Communicated by Pro- 
fessor Geikie, F.R.S. (Plates VII. to X.), . . ' .73 

V. Gaseous Spectra in Vacuum Tubes, under small Dispersion and at 
low Electric Temperature ; including an Appendix III., by 
Professor Alexander S. Herschel, M.A., Newcastle -on-Tyne. 
By Professor Piazzi Smyth, Astronomer-Royal for Scotland. 
(Plates XI. and XII.), .93 

VI. On a Special Class of Sturmians. By Professor Chrystal, . 161 



VI CONTENTS. 



PAGE 



VII. On the Cranial Osteology of Bliizodopsis. By Ramsay H. 
Traquair, M.D., F.R.S., Keeper of the Natural History Col- 
lections in the Museum of Science and Art, Edinburgh, . . 167 

VIII. On the Action of Phosphide of Sodium on Haloid Ethers and on 
the Salts of Tetrabenzyl-Phosphonium. By Professors Letts 
and N. Collie, Esq., F.R.SS. L. & E., . . . .181 

IX. On the Geology of the Fceroe Islands. By James Geikie, LL.D., 

F.R.SS. L. & E. (Plates XIII. to XVI.), . . .217 

X. Researches in Contact Electricity. By Cargill G. Knott, D.Sc. 

Communicated by Professor Tait. (Plate XVII.), . . 271 

XI. On Phosphorus- Betains. By Professor Letts. (Plate XVIIL), 285 
XII. On Dust, Fogs, and Clouds. By John Aitken, . . . 337 

XIII. The Effect of Permanent Elongation on the Specific Resistance of 

Metallic Wires. By Thomas Gray, B.Sc, Demonstrator in 
Physics and Instructor in Telegraphy, Imperial College of 
Engineering, Tokio, Japan. (Plate XVIII^.), . . . 369 

XIV. On the Histology of the Pedicillariw and the Muscles of Echinus 

sphsera (Forbes). By Patrick Geddes, F.R.S.E., Lecturer on 
Zoology in the School of Medicine, Edinburgh ; and Frank E. 
Beddard, B.A., Assistant Demonstrator of Zoology, Oxford. 
(Plates XIX. to XXL), 383 

XV. On some New Species of Fossil Scorpions from the Carboniferous 
Rocks of Scotland and the English Borders, with a Review of the 
Genera Eoscorpius and Mazonia of Messrs Meek and Worthen. 
By B. N. Peach, A.R.S.M., F.R.S.E., of the Geological Survey -. 

of Scotland. (Plates XXII. and XXIII. ), . . .39$ 

XVI. Ejfects of Strain on Electric Conductivity. By August Wit- 

kowski. Communicated by Sir William Thomson, . .413 

XVII. On the Constitution of the Lines forming the Low-Temperature 
Spectrum of Oxygen. By Piazzi Smyth, Astronomer-Royal 
for Scotland, . . . . . . .419 



Vlll 



CONTENTS. 



XXVII. The Dragon's Blood Tree of Socotra (Dracaena Cinnabari, 
Balf. fil). By Bayley Balfour, Sc.D., M.D., Regius Pro- 
fessor of Botany, University of Glasgow, 



PAGE 



619 



XXVIII. On a Red Resin from Dracaena Cinnabari {Balf fil), Socotra. 
By J. J. Dobbie, M.A., D.Sc, Assistant to the Professor of 
Chemistry, University of Glasgow, and G. G. Henderson, 
B.Sc, ........ 624 



•28 JUN 1887 




TRANSACTIONS 



OF THE 



ROYAL SOCIETY OF EDINBURGH. 

VOL. XXX. PART I.— FOR THE SESSION 1880-81. 

__ / 

CONTENTS. 



Page 
Art. I. — The Law of Extensible Minors in Determinants. By Thomas Mum, M.A., . 1 

II. — On some Transformations connecting General Determinants with Continuants. 

By Thomas Muir, M.A., ....... 5 

III. — Report on Fossil Fishes collected by the Geological Survey of Scotland in Eslcdale 
and Liddesdale. Part I. — Ganoidei. By Eamsay H. Traquair, M.D., F.R.S., 
Keeper of the Natural History Collections in the Museum of Science and Art, 
Edinburgh. (Plates I. to VL, which will be given in the next Part), . . 15 

IV. — On some neiu Crustaceans from the Lower Carboniferous Roclis of Eskdale and 
Liddesdale By B. N. Peach, A.R.S.M., F.G.S., of the Geological Survey ,f 
Scotland. Communicated by Professor Geikie, F.R S. (Plates VII. to X.), 73 

V. — Gaseous Spectra in Vacuum Tubes, under small Dispersion and at low Electa 
Temperature ; including an Appendix III., by Prof. Alexander S. Herschel, 
M.A., Newcastle-on-Tyne. By Professor Piazzi Smyth, Astronomer-Eoyal for 
Scotland. (Plates XI. and XII. ), . . . . . .93 

VI. — On a Special Class of Sturmians. By Professor Chrystal, . . .101 

VII. — On the Cranial Osteology of RhLsodopsis. By Ramsay H. Traquair, M.D., 
F.R.S., Keeper of the Natural History Collections in the Museum of Science 
and Art, Edinburgh, . . . . . . . .167 

VIII. — On the Action of Phosphide of Sodium on Haloid Ethers and on the Salts of 
Tetrabenzyl-Phosphonium. By Professor Letts and N. Collie, Esq., F.R.SS. 
L. & E, 181 

IX. — On the Geology of the Fcerde Islands By James Geikie, LL.D., F.R.SS. L. & E. 

(Plates XIII. to XVI), 217 

X. — Researches in Contact Electricity. By Caegill G. Knott, D.Sc Communicated 

by Professor Tait. ' (Plate XVII), . . . . . .271 

XL — On Phosphorus-Betaines. By Professor Letts. (Plate XVIII), . . 285 

XII. — On Dust, Fogs, and Clouds. By John Aitken, ..... 337 

(For rcmahidcr of Contents see last pa/je of Cover.) 



v^ 



/ 




TRANSACTIONS. 



I. — The Law of Extensible Minors in Determinants. 
By Thomas Muir, M.A. 

(Received 21st February 1881.) 

§ 1. As a preliminary to the establishment of the law in question, it is 
necessary to state and exemplify another law to which I have elsewhere directed 
attention, viz., 

The Law of Complementaries. # 

To every general theorem which takes the form of an identical relation between 
a number of the minors of a determinant or between the determinant itself and a 
number of its minors, there corresponds another theorem derivable from the former 
by merely substituting for every minor its cofactor in the determinant, and then 
multiplying any term by such a power of the determinant as will make the terms 
of the same degree. 

For example, taking the well-known identity employed by Hermite, 



I «A I I «2^3 I I a ih I 



ciiC 



a 2 Ca 



a,r. 



a x d. 2 \ \a 2 d. i \ |a 3 rf 4 ! 






a x a. 2 a 3 ffl 4 
I l x b. 2 b 3 b i 



Co c. 



d x d. 2 d$ d t 



a.M 



2 M '3 



• (1) 



* I do not know who was the first discoverer of this law. It presented itself to me when correct- 
ing the proof of my paper on " General Theorems in Determinants" (Trans. Eoy. Soc. Edin. 1879). 
But it must have been known to Professor Cayley before then, for in a note to a paper by Professor 
Tanner (Mess, of Math. 1878), he refers to it as a means by which Professor Tanner's corresponding 
law for Pfaffians might be established. 

VOL. XXX. PART I. A 



2 



THOMAS MUIR ON THE LAW OF 



and substituting for each determinant its complementary minor in the deter- 
minant | a 1 6 2 c 3 ^ 4 1 , we have 



c 3 d 4 | M 4 | i^d,! 
b 3 d A \ 1M4I I Mai 

I h c i I I V 4 I I M2 I 





61 6 8 & 4 


b 1 b 2 5 4 i 


= 


Cj c 3 c 4 


Cj {?2 C4 




^1 ^3 ^4 


C?! rfo f^ 4 



(2) 



a special case of a theorem of Sylvester's in regard to compound determinants. 

It is thus seen that in virtue of the Law of Complementaries the theorems 
of determinants range themselves in pairs, like pairs of theorems in geometry 
in virtue of such a law as that of Reciprocal Polars. 

$ 2. We come now to 



The Law of Extensible Minors. 

// any identical relation he established between a number of the minors of a 
determinant or between the determinant itself and a number of its minors, the 
determinants being denoted by means of their principal diagonals, then a new 
theorem is always obtainable by merely choosing a line of new letters with new 
suffixes and annexing it to the end of the diagonal of every determinant, including 
those of order 0, occurring in the identity. 

The proof is dependent upon the Law of Complementaries, and upon the 
simple fact that every minor of a given determinant is also a minor of any 
determinant of which the given determinant itself is a minor. Let (A) be 
the established identity, and | a 1 b 2 c 3 ... 4 1 the determinant whose minors are 
involved in it. Then taking the complementary of (A) with respect to 
I a-Jj 2 c 3 . . . l n \ we have an identity, (B) say, likewise involving minors of 
a 1 b 2 c 3 . . . l n \. But these minors are also minors of | a 1 b 2 c 3 . . . l„r a s fi . . . z a |, and 
therefore it is allowable to take the complementary of (B) with respect to this 
|extended determinant. Doing this we pass, not back to (A), but to a new 
theorem (A') which is seen to be derivable from (A) by annexing to the end of 
the diagonal of every determinant in it the line of letters /* ^ ■..#„,■ The law 
is thus established. 

The clause " including those of order " is necessitated by the last clause 
in the enunciation of the Law of Complementaries. 

Taking as an example the simple identity 



I «A C 3 I = «1 I &2 C 3 I — «2 I Va I + "3 I V 2 > 



and using only one new letter d and one new suffix 4, we change 



into 



EXTENSIBLE MINORS IN DETERMINANTS. 

\d 1 b f fi 3 \, «!, I&&I, « 2 , IA<M» a s , \l x e 2 \, 

a-pfod^, \a l d 4 \, b 2 e 3 d 4 1 , ■. \.a % d A \ , \b 1 c 3 d i \, \a&\, | b^ \, 



and noting that without further change the two sides would not be of the same 
degree, we annex the factor d 4 to the left hand side, thus, as it were, extending 
the process of elevation of order to an imaginary determinant of order 0. The 
result is the identity 



d t | «Ac 3 d 4 | = | a^ | | b.f^i | - | a./li | | b^ | + | a 3 d i | | b i c 2 d i 



■ (3) 



This is verified by observing that 



| a x a 2 a 3 a 4 a 4 

i 5j b 2 b 3 b i 

d i | a l b 2 c 3 d i | = | Cj c 2 c 3 c 4 

j f? x d 2 d 3 d i 

d, 



1 rtj a 2 a 3 a 4 a i 
b x b 2 b 3 \ 



c x c 2 c 3 f 4 



d-L d 2 d 3 d 4 
f?! rf„ rf 3 <f 4 rf 4 



and then expressing the last determinant in terms of products of complementary 
minors, one factor of each product being formed from the first and last line. 

Taking the identity numbered (1) above and choosing the extension e b , we 
have 

\afi 2 e b \ \a 2 b 3 e b | \a 3 b i e b \ 

I «l C /o I I «2 C 3 C 5 I I «3 C 4«5 i 

\.a x df & \ \a 2 d 3 e b \ |« 3 rf 4 e 5 | 
The corresponding extensional of (2) is 



= |MM« S | I «2<" 5 I I «3«6 I 



(4) 



\c 3 d 4 e b \ \c l d i e b \ \c x d 2 e b \ 
I'M^I \b l d i e b \\b 1 d 2 e b \ 



&i<¥*4«6 I I V 2 ^5 I «5 



(5) 



The identities (1) and (2) however admit each of two forms of Extensional, 
according as we look upon the letters in the right hand members as being mere 
elements, or as being determinants of order 1. Thus from (1) we have 



\a l b 2 e b \ \a 2 b 3 e b \ \a 3 b^e b \ 

I «!<% I I «2 C 3 e 5 I I «3^5 I 

\a x d 2 e b \ \a 2 d 3 e b \ \a 3 d i e b \ 



«i« 5 1 1^5 1 la 8 e 5 | Ke 5 | 

he b | | b 2 e b | | b 3 e b | | \e b | 

I Ci<? 5 1 | c 2 e b | | c 3 e b | | c 4 e 5 1 

'd 1 e i \ \d 2 e b \ \d 3 e b \ |d 4 e 5 | 



|« 2 e 5 | \a 3 e 



3^5 



. (6) 



and from (2) 



T. MUIR ON THE LAW OF EXTENSIBLE MINORS IN DETERMINANTS. 

I Ma I I Ma I I Ma I 
I Ma I I Ma I I Ma l 

I h<Yo I I Va I I & i c A I 
In corroboration of these, we observe that from (4) and (6) we deduce 





l&Al IVsl I&aI 




I&aI I&aI I*aI 


= 


1 <A 1 1 <A 1 1 C A 1 


X 


1 <A I | r/ 5 | | c 4 e B | 




\dfy\ l^ol |r/^| 




Irf^jsl |f?/ 5 | UVal 



(7) 



«AI l«A 

V.J I Ms I IV5I l&A 



Clfr, 



Cjfr. 



C«C« 



c.^ 



I«*aI I rf 2 «s I l^s«sl l<*A 



= I flAMA i e « 



(8) 



which is the extensional of the manifest identity 



«1 


a, 2 


a » 


a 4 


h 


h 


h 


h 


<h 


C 2 


c z 


C 4 


d x 


d. 


d. 


d, 



= ' 01W4 1 



§ 3. Thus in theory of determinants every general theorem in the form of an 
identity has its complementary and its extensional. The exact relation between 
the two latter is seen from the proof which has been given above, and may be 
formulated as follows : — If the Complementary of (A) with respect to a certain 
determinant be (B), its Complementary with respect to a determinant of higher 
order is the Extensional of (B). Consequently, if, as sometimes happens, the 
Complementary of (A) with respect to a certain determinant be (A) itself, its 
Complementary with respect to a determinant of higher order is its Extensional. 

By the two laws the theorems of determinants are knit together in a way 
which is interesting theoretically, and which at the same time has the practical 
advantage of making the remembrance of the whole body of theorems a very 
simple matter. 



( -5 ) 



II. — On some Transformations connecting General Determinants ivitk 
Continuants. By Thomas Muir, M.A. 

(Read 21st February 1581.) 

§ 1. It is well known that by a simple transformation of a determinant we 
may cause a zero to take the place of any one of the elements. The theorems 
of Hermite * and HoRNER,t for example, for depressing the order of a determi- 
nant may each of them be viewed as the result of repeated transformations of 
this kind, the operation being continued until all the elements of a row or 
column except one are replaced by zeros. 

With these facts in view, it occurred to me about a year ago to test the 
possibility of transforming a general determinant so as to have zeros in every 
one of the positions held by them in Sylvester's continued-fraction determinant, 
viz., everywhere except in the principal diagonal and the two bordering minor 
diagonals. The transformations to which I was then led form the subject of 
the present short paper. 

§ 2. Beginning with the determinant of the fourth order | a^c^d^ | we have 
as the result of a first transformation 



I «A C 3^4 I = 



Ift^l \a 2 d 3 \ a 3 a i 

\b x d 2 \ \b 2 d 3 \ b 3 \ 

\c x d 2 \ \c 2 d 3 \ c 3 c 4 

d 3 d 4 



+ *A . 



and multiplying each element of the first column here by | c 2 d z | and diminishing 
the result by | c x d 2 | times the corresponding element of the second column, we 
have 

d 2 \a x c 2 d 3 \ \a 2 d 3 \ a 3 « 4 



I « A C .A I = 



d 2 \b x c 2 d 3 \ \b 2 d 3 \ b 3 Z> 4 
\c 2 d 3 \ c 3 c 4 




d 3 d± 



™ tvottql Vftl'l 



where on one side of the principal diagonal the resulting determinant is of 



* Liouville's Journal, xiv. p. 26. 

t Quart. Journ. of Math., viii. pp. 157-162. 



VOL. XXX. PART I. 



6 



THOMAS MUIR ON SOME TRANSFORMATIONS CONNECTING 



the form desired. Operating in an exactly similar way on the other side, we at 
once obtain the kind of result which was hoped for, viz., 



(l) 



\a 1 \c z d i \ = 



1 <hpJs 1 


1 Ms 1 ° 





1 V 2 ^3 1 


1 Ms 1 1 «A 1 








1 C A 1 1 «2 C 3 1 


1 Ms C 4 1 





|« 2 tf 3 | 


1 « 2 M 4 1 



-H«Al I M 3 1 M 3 I> 



the non-zero elements making their appearance as determinants in virtue of the 
well-known theorem 






=i A^o.y^ y \ 



For the case of the determinant of the fifth order the corresponding identity is 



(!') 



\*jf>&flfy\=- 



IftjC/^l iMj^l 

I Vsdjft I I M 3 e 4 I I «2 & 3 g 4 I 

|c 2 d 3 e 4 | |« 2 c 3 e 4 | |« 2 V 4 I ° 

| a/? 3 e 4 1 | a 2 \d i 1 | a 2 & 3 c 4 d 5 1 

| a 2 & 3 e 4 1 | a.fi.^e^ | 



|« 2 & 3 c 4 | |a 2 & 3 e 4 | |« 2 d 3 e 4 | |c 2 d 3 e 4 | 



the general law of formation of the right-hand member being contained in the 
following rule : — To obtain the first part, viz., the continuant, take the original 
determinant | a-J}^ . . . . z n \; from the first column delete the elements which 
in a continuant are zeros, and replace them by zeros, writing all the deleted 
letters in order alongside each of the remaining elements of the column ; treat 
the other columns in the same way ; affix such suffixes to these added letters 
that the suffixes of the first column may be 1, 2, 3, . . . , n— 1, of the last column 
2, 3, 4, ... , n, and of each of the intervening columns 2, 3, 4, . . . , n — 1 ; enclose 
each set of suffixed letters in determinant brackets. To obtain the second 
part, viz., the divisor, take the product of all the elements of the continuant 
which border its principal diagonal, excepting those in the first and last columns, 
and rejecting duplicates. 



§ 3. Let us now return to our first determinant | « 1 J 2 c 3 6? 4 1 . Multiplying each 
element of the fourth column by m z and diminishing the result by ?n 4 times the 
corresponding element of the third column, and treating the third and second 
columns in a similar manner, we have 



GENERAL DETERMINANTS WITH CONTINUANTS. 



| a x b 2 c 3 d i | = 



a x Im^l 1^2*3! Iw 3 a 4 | 

b x \m x b 2 \ |m 2 & 3 | |w 3 J 4 | 

e, I wi^ I I m. 2 c 3 1 I m 3 c 4 1 

d) I m^ 1 I m 2 d 3 1 I ??i 3 rf 4 1 



■m 3 ?« 2 mj 



and subjecting this new determinant to the set of operations to which | a 1 b 2 c B d i 
itself was subjected in § 2, we finally obtain 



l«l&2 C 3^J =- 



I fl^dg 1 I m 1 « 2 rf 3 1 

\b x c 2 d 3 \ Im^jl Iw^&sl 

I m^ds I I wi^a, | | m l a 2 bc i c i \ 

\ , m l a i d z \ \m x aJ)gL^ 



Imjfl^sl lwh«2^l l m i c 2 f ^ 



(2.) 



A comparison of this with (1) brings out the fact that the right-hand member 
there is not altered if we change the elements of the last three columns, and 
the factors of the divisor, all into determinants of the third order by inserting 
an m x in each. 



§ 4. Making use now of the Law of Complementaries we return to (1), and 
substitute for each determinant its complementary minor in | a-p.f^d^ \ . This 
gives us the new identity 



I a A I I & i c 4 1 I M 4 1 = 



a relation connecting only the elements 



b 4 |VJ 

a 4 IffjcJ U^J 

\a 1 b i \ IMil <*i 

I b x c 4 I c x 



(3) 



« 1 J 1 c 1 </ 1 
a i b 4 c 4 d i , 



so that as there are six pairs of these lines, we can at once write six 
identities like (3), and thence find by the Law of Complementaries six 
identities like (1). 



8 THOMAS MUIR ON SOME TRANSFORMATIONS CONNECTING 

The corresponding relation for two rows of Jive elements is 



I «A I I V 5 1 I C A I I f V 5 1 = 



h 


l&Al 











a 6 


1 <h°5 | 


1 <v*5 1 











1 «A 1 


\h(h\ 


1 d & 1 











Ital 


1 Cl«5 1 


H 











1 c x ^ 5 1 


h 



(3') 



unci from this and (3) the general theorem is apparent. 

Again, taking the complementary of (2) with respect to | m x a^> z c^d h \ , and, 
for the sake of comparison with (3), changing the suffixes 4, 5 into 1, 4 respec- 
tively, we have 



m 4 a-fii | b 1 c i | | Cjd 4 | = 



\m l l i \ \lyCi | 

\m 1 a i \ la^ | | c x d 4 1 

K&4I 1 5^ I d 4 

IV4I C 4 



(4) 



Avhere the right-hand member differs from that of (3) only in the first and last 
columns. 



§ 5. By taking the complementary of (3) with respect to | a^b.f^d^ | we should 
of course return to (1) ; by taking the complementary, however, with respect 
to J a-fi^d^e^ I (or | a^ 2 c 3 d 4 e 5 f 6 . . . | ) we obtain a new result, viz., 



a-Jj.^d^ 





«1^3 e 5 1 


a 2 d 3 e 5 1 










1 ^2^5 1 


1 Ms e 5 1 


\a 2 b 3 c 5 \ 










1 c 2 d 3 <% 1 


\ ci 2 c 3 e 5 1 


1 «A C 4*5 1 










1 a 2^3 6 5 ' 


1 ^Ma 1 



a 2 b 3 e b \ \a 2 d 3 e 5 \ \c./l 3 e b 



<5) 



which it is interesting to compare with both (1) and (!'). 



§ 6. If in (2) we write b, c, d, e for a, b, <:, d respectively, a for m, and 2, 3, 4, 5 
for 1, 2, 3, 4 respectively, the left-hand member of (2) will be a principal minor 
of the left-hand member of (!') and the first determinant on the right-hand of 



GENERAL DETERMINANTS WITH CONTINUANTS. 9 

(2) will be the corresponding minor of the first determinant on the right-hand 
of (1') ; and from the two identities we shall have by division 



| «jC 2 f?3e 4 | | « 2 ^3 e 4 I 

IVs^J \\d^\ \a 2 i 3 e 4 \ ■ 

| c 2 d 3 e 4 | | « 2 c 3 6 4 1 | « 2 & 3 c 4 | 

K^Vil \ a A d i\ I«2&3 C AI 

0| « 2 & 3 e 4 1 I « 2 5 3 c 4 c 5 1 



\a l b 2 c 3 d i e b \ \c 2 d z e i 



| & 2 c 3 tf 4 e 5 1 



I c 2 d 3 e 4 1 | « 2 c 3 e 4 I I a A r i I 

|M 3 e 4 | |rt 2 M 4 l |«2 & 8 C AI 

| « 2 &3 e 4 I I «2 & 3 C 4 e 5 I 



Changing the numerator and denominator on the right hand in accordance with 
the theorem of which 



a d 




a 


-df 


/ b e 


= 


-1 


h -erj 


g c 







-1 c 



is an example, we have by Sylvester's fundamental theorem regarding the ap- 
plication of continuants * 



\a J b 2 c. i d i e b \ \e. 2 d z e i \ 
I hcA e 5 1 



= | a x c 2 d z e,± | 



\a 2 d. i e i \ l&xCa^l 



\hJUA l«A!iM¥^ 



(6) 



o 2 c 3 e 4 |— ■ 



q. 2 & 3 e 4 | k 2 ^ 3 e 4 | 
12341 | a 2 Ws | 



The corresponding identity for determinants of the next lower order is 



\a 1 b,c. i d i \ | 


c 2 d 3 \ 




a x c 2 d 3 1 


1 a A 1 
i M 3 1 


1 V 2 ^3 1 
1 «2&3 1 

l«2 C 3|- 


1 c 2 d 3 1 

l«2 


& 3 c 4 | | a 




1 \h d i 


1 


2 f 4l 






1 « 2 M 4 





(6') 



The general theorem is readily formulated by attending to the rule given in § 2 
for forming the continuant in (1) and (1'). 

§ 7. If in (4) we write b, c, d, e for a, b, c, d respectively, a for m, and 5, 1 



* Phil. Mag., 4th ser., vol. v. pp. 446-456. 



10 



THOMAS MUIR ON SOME TRANSFORMATIONS CONNECTING 



for 1, 4 respectively, and then proceed with (3 7 ) and (4), as we have just done 
with (!') and (2), we find 



I«AI _, 



fl.c 



1^51 






l^Al IVs 



(V) 



c,c 



1^5 I 






This is the complementary of (6), and might so have been obtained. As it is 
easily verified, we can therefore readily have by means of it a verification of the 
more important theorem with which it is related. 

§ 8. In the continuants of §§ 2-5 the zeros were introduced by operating- 
only with rows upon rows, or with columns upon columns. If now, however, 
we introduce those on the one side of the principal diagonal by operating with 
rows upon rows, and those on the other side by operating with columns upon 
columns, we obtain a result quite distinct in character, and not less interesting, 
viz., we have 



where 





Vi 


x x 















z i 


ty 


x 2 















H 


Vz 


*8 















*3 


Vi 


x i 




n ,h c^rl '. f 1 — 











z i 


y 5 




a 


2 1 ^2^3 1 


1 «2&3 C 4 


1*1 


\b x c 2 


1 1 &iC a C? 3 1 


Xy=-d^ , 








Z x - 


h. 


x 2 — | a 2 b 3 1 , 








2 2 = 


=1 V 2 I. 


x 3 —b x |a 2 & 3 c 4 |, 








«3= 


:a. 2 {b&dsl, 


x^—\b x c 2 \ \a.,b 3 


Wh 1 > 






*i = 


:| a 2 b 3 1 


I b x c 2 a 



(3) 



and 

?h=h > 

y 3 =a 2 \b x c 3 \ — a 3 \b x c 2 \, 

y 4 = I a 2 b 3 I I b x c 2 d 4 \ — | a 2 b 4 \ \ b x c 2 d 3 \ , 

y r = \ « 2 & 3 c 4 1 I b x c 2 d 3 f 5 1 - 1 a 2 b 3 c 5 | | b^d^ | . 

The process of transformation is not given, because to do so would unneces- 
sarily lengthen the paper ; the reader, however, will find it worthy of attention, 
one or two little-known identities turning up in the course of it. 

The corresponding expression for | a 1 b 2 n z d i | is got by merely deleting the 
last row and column of the numerator, and the last two factors of the deno- 
minator. 



GENERAL DETERMINANTS WITH CONTINUANTS. 



11 



The theorem related to (8), as (2), or rather that form of (2) used in § 6, is 
related to (1'), is 



hhdiA\ = 





y% 


*2 












z 2 


y 3 


X 3 












«3 


Vi 


x t 












h 


y$ 





«2 l«2& 3 | | «2 & 3 C 4 I h \h C 2\ IV^sl 



(9) 



where x 2 , y 2 , z 2 , . . . have the same signification as before. This may be obtained 
after the manner of (8), but (8) having been proved, (9) at once follows as the 
result of differentiation with respect to a 1 . 
From (8) and (9), by division, there comes 



aj) 2 c % clj h 



b 2 c 3 c hf 5 1 



?h- 



X-iZ- 



1*1 



Vi - 



x ft-2 

its 



(10) 



x»z> 



3*3 



Vi 



2/5 



an identity more notable than those of like kind previously given. 

§ 9. The result of taking the complementary of (8) is peculiar, the left-hand 
member remaining unchanged. We thus obtain still another expression for 
| a-ficf^d^ | , which would not readily have been lit upon otherwise, viz., 





*h fi 









(fl »?2 ?2 









£2 % 


& 






r 3 


^ & 




\rt h r rl f \ — 





S* ^5 




1 a i u -2h a iJ5 1 — 

K/ 5 I K/al IM4/5I IM4/5 

where 


1 <W*4/ 3 1 1 W*/a 1 


& =M 4 / 5 | 1«i*iA/5j1j 


?2 = I«3^/ 3 I, 


%3=\a-2 c Af-o\ \ d iA\, 


f3=[5 1 e 3 ^ 4 / 5 | |« 4 / s |, 


and 


C=K'l^/ 5 l «5, 


Vi=Wc%difi,\ , 




% = I«1 C 3^/ 5 I> 




% = l & 1 C 3^/ 5 I IM4/5H 


1 Ma^/s 1 1 «A/ 5 1 > 


*?4 = l C A/j>l l« 3 / 5 |-| C l 


*b/bI 1 "4/5!. 


*?5 


=1 ^1/5 k- 1^1/4 1«5- 







(11) 



12 



THOMAS MUIR ON SOME TRANSFORMATIONS CONNECTING 



Taking the complementary of (9) we have 



% 


6 








r 2 


% 


6 








& 


Vi 


6 








£4 


% 



l rf l/ 6 l I tt 4/5 I M4/5I 1^4/6 I IW^/sl I61M4/5 

and thus again, by division, there comes 



(12) 



I «AC 3 ^ 4 /5 I _ 



>?I 



V2 



U, 



250 



(13) 



% 



Ml 

Vi 



V 5 



— the complementary of (10). 

If the values of the £'s and £'s be compared, it will be seen that there is 
something abnormal in the second line. This is not due to an error ; the 
factor I « 1 J 2 c 3 6? 4 / 5 1 must appear in one of the two elements £ 2 , £ 2 > and ma y 
appear in either, but not in both. 

§ 10. In (10) we have a continued fraction found as an expression for the 
quotient of a determinant by a differential coefficient of it with respect to one 
of the elements. This was obtained from the two distinct theorems, (8) and 
(9), by division, &c. Owing, however, to a peculiarity of (8), we do not need 
the assistance of (9) to obtain such a result. Taking, instead, the identity 
corresponding to (8) for the case of the determinant j a l b 2 c 3 d i | , we have from 
it and (8), by division, 



a AcAf b \ K&3C4! \hh d A 



a 1 &. 2 c 3 ^4 1 



Hi x \ 



2/2 

*2 























vCo 







Vi 


tXs^ 




Z 4 


y& 


1 



V\ %1 



«2 







x 2 

3/3 X 3 



and, continuants being unaltered in substance by having the order of the 
elements in their diagonals reversed, there thus results 



<h!>#Afb 



a 2 h c i 



l x c 2 d z 



« A C 3^4 



X i Z 4 


XoZ 3 


x 2 z 2 


• 


Vi ' 


?h 


x \ z l 



(14) 



V\ 



There is evidently no theorem corresponding to (6) or (7), as this corre- 
sponds to (10), the continuants employed in finding the former having a 
.symmetry with respect to both diagonals. 



GENERAL DETERMINANTS WITH CONTINUANTS. 



13 



Note on Mr Muir's Transformation of a Determinant into a Continuant. 

By Professor Chrystal. 

(Read 21st February 1881.) 

I. The following way of arriving at some of Mr Muir's elegant theorems may be of some 
interest : — 

Consider the system of equations, 



(nK+(i2)3 8 + ... +(i»K=o, 

(21>» 1 + (22)a>2+ ... +(2m)x„ = 0, 
(n\)x x + (n2)x., + . . . + (nn)x n = 1 , 



(1), 



the left hand sides being zero in all but the last. Let A be the determinant of this system. 

From the first n - 1 of these equations we can eliminate all the variables but x\ and ar 2 in one way ; 
and all but x,._ u x r , and x r+l in n - 1 ways ; also from all the n equations {i.e., from any n - 1 of them, 
the last being always included) all the variables but #„_ a and x„ in n - 1 ways. We thus get 



EjA-j + F^ 



= 



~l\..r r ^ + E r x r + F r x r+1 =0 



i 
Where the determinants D, E, F are derived from A as follows, by omitting 



(2). 



In 


Columns. 


Rows. 


1 


E, 


1st 






D,. 

E, 

E, 




s' K and u tk 
s u and n ih 
s'* and »'* 


s is any number 
common to these 
three except n 


>•'* and r-l'f' 


r-l]'*and r+lf 
r* and r - 1|"' 


E. 
F„ 


»" 


t"' 

r h 

t <h and a'* 


t is any number 
common to these 
three except n 


»-l|«* 


«* and n~l|* 



VOL. XXX. PART I. 



1 4 TRANSFORMATION OF A ■ DETERMINANT INTO A CONTINUANT. 

Solving now the second set of equations, we get 

.-•, Ej F, I = . F t . ■ ■ I ' 

D 2 E 2 F 2 ; . j. 0,E 2 F 2 

D 3 E 3 F 3 D 3 E 3 F 3 



I*., i E„ , F„_i ' 

D„ E„ F„ 0. 



-(-i)^r 1 .,..F ll . 1 F, i 



but from the first set 
Hence 



X 


i* = 


■(" 


1)' 


*i- 


*1 


F 2 


F 


. . . 


• F„ 


E x 


E 2 






■ F M 


D 2 


D 3 




y» 


■ D„ 


F 2 


F S 




- F < 



(3); 



■whore it is to be noticed that every row of the continuant except the first contains an index susceptible 
of ri - 1 different values. 

By giving these indices the proper values, we get, as particular cases, Mr Muir's formula? 
(1). (l')and(5). 

By solving for x„ , we get at once a result like that of Mr Muir in § 6. 

II. The above may be generalised as follows : — 

Noticing that every one of the equations in set (2), except the first, is susceptible of n - 1 different 
forms, and multiplying each of these by one of the arbitrary quantities m rl m r2 . . . m rn _ lf we get, by 
addition, in each case a new equation. Hence (writing also for uniformity's sake e^f x for E t and Fj), 
we get the new set 

e x Xy +f x x a =0 

d,.r r i + e r x r +/ r x r+1 =0 } (4). 

^n*n-l + «»«»=/„ 

"Whence, as before, 

if /,.../« v 

A = l e % e J . . . (5); 

fj, ■ ■ /„ h • ■ 

where e l f 1 are the same as E T F, ; but d r e T f r now omit only the ri* row, and have each an additional 
first column ?/>, , m r ... m rn . . 

It is to be observed that in each of the rows of the continuant we have a different set of n - 1 
arbitrary quantities. The identity (5) is therefore one of considerable generality. ! It gives Mr Muir's 
identity (2) as a particular case. 



( 15 ) 



III. — Report on Fossil Fishes collected by the Geological Survey of Scotland 
in Eskdale and Liddesdale. Part I. — Ganoidei. By Ramsay H. 
Traquair, M.D., F.R.S., Keeper of the Natural History Collection 
in the Museum of Science and Art, Edinburgh. (Plates I. -VI.) 

(Read 19th July 1880.) 

INTRODUCTION. 

I am indebted to the kindness of Professor Ramsay, Director-General, and 
of Professor Geikie, Director of the Scottish Branch of the Geological Survey 
of Great Britain, for the privilege of examining and describing a remarkable 
collection of fossil fish-remains from the Lower Carboniferous rocks (Calci- 
ferous Sandstone Series) of Eskdale and Liddesdale. Most of the specimens 
were collected by Mr Arthur Macconochie, one of the collectors attached to 
the Scottish Geological Survey ; and Mr Walter Park of Brooklyn Cottage, 
Langholm, has also willingly co-operated in the search, so far as the district of 
Eskdale is concerned. I have myself also had the pleasure of twice visiting 
Eskdale, along with Mr Macconochie and Mr B. N. Peach, and on these 
occasions I obtained a few specimens for my own collection. 

This collection is of the greatest possible interest, both from a geological 
and from a zoological point of view — both as opening up to us an almost 
entirely new Scottish Carboniferous fish-fauna, as well as from the purely 
zoological interest attaching to the structural peculiarities of many of the new 
forms themselves. My own business with these fossils is, of course, entirely as 
a zoologist. 

The fish-remains which have occurred in these strata are referable to the 
orders of Ganoidei, Dipnoi, and Selachii, of which only the first will be consi- 
dered in this instalment of the report, while a second part will be devoted to 
the enumeration and description of those belonging to the two remaining 
Orders. 

The following is a list of the genera and species of Ganoids which have 

VOL. XXX. PART I. D 



1(J 



RAMSAY H. TRAQTJAIR's 



occurred, an asterisk being prefixed to the names of those species which are 
here described as new : — 



9 
10 
11 
12 
13 
14 
15 
16 
17 
18 
19 
20 
21 
22 



23 
24 
25 
26 

27 



28 



ACANTHODID^E. 

Acanthodes, sp. . 

Ehizodontid^e. 

Strepsodus sauroides, Binney, sp. 
Archichthys Portlockii, Ag. 

Saurodipterid^e. 
Megalickthys, sp. 

CCELACANTHID/E. 

Ccelacanthus Upturns, Ag. 

* Huxleyi, Traq. 

PaljEoniscid^:. 

* Elonichthys serratus, Traq. 

* pidchcrrimus, Traq. 
Rhadinichthys Geihiei, Traq. 

* dclicatulus 

* Macconochii, Traq 

* tubermlatus, Traq 

* miyusttdus, Traq. 

* fusiformis, Traq. 
*Cycloptychius concentricus, Traq 
* Phanerosteon mirabilc, Traq. 
*PTolurics Parld, Traq. 

* fulcratus, Traq. 
*Canobius Bamsayi, Traq. 

* elegantidus, Traq. 

* pidchcllus, Traq. 

* politus, Traq. 

Pla.tysomid.4<:. 

Eurynotus crenatus, Ag. 

* ? aprion, Traq. 
Wardichthys ? cyclosoma, Traq. 

* Cheirodopsis Geikici, Traq. 
* Platysomus superbus, Traq. 

TARRASIID.E. 

*Tarrasius prdblematicus, Traq. 



Eskdale. 



+ 



+ 
? 



+ 



+ 



+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 
+ 



+ 
+ 
+ 



+ 



Liddesdale. 



+ 
+ 



+ 



+ 
+ 
+ 



REPORT ON FOSSIL FISHES. 17 

Concerning the above list, there are three things which principally strike 
the attention, — 

1. The occurrence of a large number of forms perfectly new to science. I 
have endeavoured most strenuously to avoid all undue multiplication of genera 
and species ; indeed, I may have erred in the opposite direction ; yet out of 
twenty-eight species of Ganoids occurring in these beds, at least twenty 
must be described as previously unknown. Of fourteen genera, five are new, 
namely, Phanerostecm, Holuras, Canobius, Cheirodopsis, and Tarrasius, while 
the last named genus is altogether so peculiar that I can find no place for it 
in any known family. Some amount of change in our notions of the definition 
and limits of the family Paheonisciclse will also be necessary, if the genera 
Holurus, Phaner -osteon, and Canobius are to remain where I have placed 
them. 

2. The absence or paucity of forms characteristic of rocks of similar age on 
the northern side of the southern uplands of Scotland. There are no remains 
which can with certainty be referred to the genus Rhizodus, which in central 
Scotland occurs abundantly from the bottom of the cement-stone group upward 
through the Carboniferous Limestone series. The well-known Eurynotus 
crenatus of Mid-Lothian and Fifeshire is represented only by a few scales and 
bones from Liddisdale. And as regards the Palseoniscidae, all are new save 
one, which I refer, not without doubt, to Rhadinichthys Geikiei, a species de- 
scribed by myself in 1877 from the Wardie shales of Colinton, near Edinburgh. 
Even the characteristic Cement-stone and Edge-coal type of the genus Elon- 
ichthys, that of EloniclUhys Robisoni, is represented only by one rare species, 
Elonichthys serratus, and that also new. 

3. The passing down into the Calciferous Sandstone Series of genera, 
hitherto known as characteristic of the Coal Measures or Upper Carboniferous 
series of rocks, although most of these have, it is true, occurred sparingly in 
the Carboniferous Limestone series. Strepsodus, Coelacanthus, and Platysomus 
are best known to us as Coal Measure genera ; and although fragmentary 
remains of them have been found also in rocks of the Scottish Carboniferous 
Limestone series, their appearance in the subjacent Calciferous Sandstones is 
now observed for the first time, while Cycloptychius has not hitherto occurred 
in any horizon below the Millstone Grit. 



18 RAMSAY H. TRAQUAIR'S 

DESCRIPTION OF GENERA AND SPECIES. 

Order GANOIDEI. 

Suborder Acanthodei. 

Family AcanthodidtE. 

Genus Acanthodes, Agassiz, 1833. 

(Agassiz, Poissons Fossiles, vol. ii. p. 19.) 

Several imperfect specimens of Acanthodes have occurred in the Eskdale 
beds, but in the present unsatisfactory state of our knowledge of the British 
Carboniferous members of this genus, it may be safer to leave them for the 
present undetermined as to species. It is to be hoped that ere long the 
accumulation of more material, from various, horizons and localities, will render 
practicable a satisfactory revision of the Acanthodidge of the Carboniferous 
formation generally. Meanwhile, the want of sufficiently definite characters 
for the species already named renders the determination of specimens, 
especially when in a fragmentary condition, a matter of extreme doubt and 
uncertainty. 

Suborder Crossopterygii. 

Family Rhizodontid^e. 

Genus Strejisodus (Huxley), Young, 1866. 

A tooth undistinguishable from those of Strepsodus sauroides, Binney 
sp., has occurred at Tweeden Burn in Liddisdale, and another at Glencartholm 
in Eskdale. To Strepsodus may also be referred some large thin cycloidal 
scales from Glencartholm, one of which measures If inch in diameter. 

Strepsodus is of rare occurrence below the horizon of the Millstone Grit, 
and the present specimens occur lower down in the series than any which have 
hitherto been found. The remains of Strepsodus sauroides constitute, as is well 
known, abundant and characteristic fossils in the bituminous shales and 
cannel coals of the True Coal Measures both in England and Scotland. 

Genus ArcMclithys, Hancock and Atthey, 1871. 
Archichthys Portlockii, Ag. sp. 

Holoptychius Portlockii, Agassiz, Poissons Fossiles, vol. i. pt. xxxvi. — name only. 
„ „ Portlock, Geol. Rep. p. 464, pi. xiii. figs. 5-11. 

M'Coy, in his " British Palaeozoic Fossils," p. 613, states that he is quite 
certain " that the Holoptychius Portlockii (Ag.) of the fish beds at Cultra 
Holywood, near Belfast, and Draperstown, &c, are identical in all characters, 
both of the teeth and scales, with the Holoptychius Hibberti (Rhizodus) of 
the Burdiehouse and Gilmerton beds." However, in his list of synonyms of 



REPORT ON FOSSIL FISHES. 19 

Rhizodus Hibberti (ib. p. 612), he has given Holoptychius Portlockii with an 
appended query. 

But an examination of the originals of Col. Portlock's figures from Mag- 
hera, Deny, now in the Museum of Practical Geology, Jermyn Street, reveals 
the unquestionable fact that they are not specifically identical with Rhizodus 
Hibberti, nor even generically, if the usual ideas as to the definition of Rhizodus 
are to be retained. The teeth are rounded or oval in transverse section, and 
devoid of the cutting edges characteristic of Rhizodus ; the folds of the base 
are proportionally large and coarse, and the surface is covered with close and 
minute yet sharp vertical striae, which fade away towards the apex as well as 
along the anterior aspect of the tooth. Now these are precisely the main 
external characters of the teeth from the Coal Measures first named by 
Messrs Hancock and Atthey Archichthys sulcidens, * and accordingly I have 
already (Proc. Roy. Soc. Edin. ix. 1878, p. 657) proposed to transfer 
" Portlockii" to the last named genus. 

From Tweeden Burn, Liddisdale, there are two teeth which I cannot 
distinguish in essential characters from those of Archichthys Portlockii of the 
Irish Lower Carboniferous rocks. Of these the larger is broken, both at base 
and apex, but when entire would I think have measured 1| inches in height. 
The transverse section is rounded ; the base displays remains of coarse 
plications ; the greater part of the surface as shown in the specimen is smooth, 
but the posterior aspect is strongly marked with the characteristic delicate 
striation. The other is half an inch in height by ^ inch in long diameter at 
the base, which also shows remains of coarse plications ; the transverse section 
is rounded, and the characteristic striae extend over a proportionally larger 
extent of the surface than in the larger specimen. In fact, this smaller tooth is 
nearly the exact counterpart of one from Maghera in the Jermyn Street 
collection. 

Associated with these teeth are numerous thin rounded scales, which 
probably belonged to the same fish, and which also, though smaller in size, 
closely resemble those of Archichthys Portlockii from Maghera. All of these 
have the outer surface attached to the matrix, and concealed, the inferior surface 
alone being exposed. But one scale from Tarras Water, Eskdale, shows some 
patches of the outer surface, and this is covered with minute granules arranged 
in closely set lines radiating from the centre. This scale I am also inclined to 
refer to Archichthys Portlockii. 

I have as yet seen no evidence that any of the Rhizodont scales from either 
Liddesdale or Eskdale belong to Rhizodus, and undoubtedly no tooth referable 
to that genus has occurred. 

* "Note on an undescribed Fossil Fish from the Newsham Coal-shale near Ne\vcastle-on-Tyne. " — 
Nat. Hist. Trans. Northumb. and Durham, vol. iv. 1871, pp. 199-201. 



20 RAMSAY H. TRAQUAIR'S 

Family Saurodipterid/E. 
Genus Megalichthys, Agassiz. 

(Agassiz, Poissons Fossiles, vol. ii. part 2, p. 89.) 

Fragmentary remains, consisting both of scales and head plates, referable 
to the genus Megalichthys, have occurred both at Tweeden Burn and near 
Glencartholm. They are far too imperfect to justify any specific deter- 
mination. 

Family Ccelacanthid.e. 
Genus Cmlacanthus, Agassiz, 1843. 

(Agassiz, Poissons Fossiles, vol. ii. pt. 2, p. 170 j Huxley, Dec. Geol. Survey xii. 1866, p. 8.) 

Cvelacanthus leptwus, Agassiz. 

A few scales which I cannot distinguish from those of the common 
Cvelacanthus Upturns of the Coal Measures (Huxley, op. cit. pi, ii. figs. 3 and 4, 
pi. iii. fig. la), occur upon a bit of shale from Tweeden Burn, along with scales 
of Earynotus and Megalichthys. 

Cvelacanthus Huxleyi, sp. nov. Traquair. 
PI. I. figs. 1-4. 

Among the specimens from Glencartholm, Eskdale, are five of a 
Cvelacanthus, which seems undoubtedly different from any hitherto described 
member of the genus. They are hardly so perfect as most of the smaller fishes 
from these rocks, and it will therefore be advisable to notice each of the 
specimens in succession. 

No. 1 (fig. 2) is the most perfect as regards figure, but its details are consider- 
ably obscured by a thin and utterly irremovable layer of matrix which adheres 
to its surface. It shows a small fish 1\ inch in length, and \ inch in depth at 
the first dorsal fin. The head, concerning whose structure nothing more can be 
said, save that it displays the general contour of that of a Cvelacanthus, is con- 
tained nearly four times in the total, the terminal appendage, or secondary caudal 
fin, being however absent. Both dorsal fins, as well as the principal caudal, are 
tolerably plainly exhibited, all showing the form and structure characteristic of 
the genus, but of the other fins there are no certain traces; and as to scales, only 
the merest " shadows," as it were, can be distinguished. The internal skeleton 
is also dimly visible, and from the crowded appearance of the vetebral spines it 
seems possible that the short stumpy contour of the fish is to some extent due 
to post-mortem shortening up. Interesting as this little specimen is, it is 
hardly possible to found much upon it beyond the generic diagnosis. 



.REPORT ON FOSSIL FISHES. 21 

No. 2 is the posterior part of a fish 1£ inch in length, being cut off just 
in front of the posterior dorsal fin, of which feeble traces are present. Behind 
this, the caudal vertebral apophyses, interspinous bones, and fin-rays are shown 
in a good state of preservation. Eighteen rays may be counted in the upper 
part of the caudal fin ; not so many are seen in the lower part, but they 
are evidently not all preserved. No scales are shown. 

No. 3 is a portion of a fish, deficient in all the fins save some traces of 
the anterior dorsal and of the ventral, and in all probability belonging originally 
to a specimen of about the same size as No. 2. The scales are pretty well 
shown, and these seem proportionally larger than in Ccelacanthus Upturns; 
they are ornamented with delicate ridges, which are proportionally fewer and 
wider apart than in the last named species, although they follow the same 
general arrangement in converging towards the middle line of the scale. The 
head is considerably crushed ; two bones are, however, very distinctly seen, 
and demand special attention. One of these is the right operculum dis- 
located from its place, and thrust away to a position close above the cranial 
shield. It has the usual trigonal shape of the operculum of Ccelacanthus, but 
what is undoubtedly its external surface is quite smooth, and devoid of the close- 
ridged ornament found in Ccelacanthus lepturus. The other bone (fig. 3) is that 
which in Ccelacanthus lepturus has been determined by Messrs Hancock and 
Atthey as the largely developed angular element of the mandible, and this, 
instead of the fine close thread-like striation of the same bone in that species, 
is marked on its outer surface only by four slightly oblique, comparatively 
coarse, and distant prominent ridges. 

No. 4 (fig. 1) is a very pretty specimen, showing as it does the vertebral 
apophyses and the remains of the ossified air-bladder with great distinctness. 
It is 3 inches in length, of which the head occupies § inch. The bones 
of the cranial roof display the same smoothness on their outer surfaces as the 
operculum in No. 3, and some coarse ridges are seen on a fragment of the 
angular element of the mandible. Beneath the lower jaw is seen the impres- 
sion of the internal surface of a jugular plate of a narrow form, its length 
being \ an inch, and its breadth hardly more than ^%. In the body twenty- 
four neural spines, bifurcated proximally, may be counted as far as the 
commencement of the haemal spines ; the bones of the caudal region are 
somewhat disturbed. Below the abdominal part of the vertebral axis the 
air-bladder is indicated by a black shining film ; the scales agree in character 
with those of No. 3. The first dorsal, the principal caudal, and some remains 
of both ventrals, as well as of the pelvic bones, are exhibited, but do not call 
for special remark. 

No. 5 is the somewhat distorted anterior part of a specimen which must 
have originally been of a larger size than any of the foregoing. It is chiefly 



22 RAMSAY H. TRAQU AIR'S 

remarkable for the clear and distinct manner in which the scale-markings are 
exhibited, and these consist of the same fine and comparatively distant ridges 
seen in the other specimens. Scarcely anything of the head remains, but a 
part of the anterior dorsal fin is present, its rays being, as usual, articulated 
towards their terminations. 

Remarks. — As assuredly the above described Ccelacanthus cannot be 
identified with any Permian or Secondary species, and as the American 
Carboniferous species seem to be closely allied to Ccelacanthus lepturus, Ag., 
it is only necessary to compare it with the latter, and with Coelacanthus 
Phillipsii, Ag., and Coelacanthus elongatus, Huxley. Of these Coelacanthus 
Phillipsii is founded upon a large tail from the Carboniferous rocks of Halifax, 
Yorkshire, and is well distinguished by its large rounded scales/' Coelacanthus 
elongatus, from the Coal Measures of Ballyhedy, County Cork, Ireland, is 
described by Professor Huxley as having a more elongated head than the 
other species, and the impressions of the bones of the skull present " traces of 
a minutely granular or lineated sculpture, "t Coelacanthus lepturus, whose 
characters, external and internal, are best known to us, have the exposed 
surface of its scales extremely closely striated, while the external cranial bones 
are everywhere covered by a very well-marked ornamentation, consisting 
of close, fine, yet sharply defined wavy and tortuous ridges and granules, which 
we search for in vain on the skull of Coelacanthus Huxleyi, where, on the other 
hand, the head bones are mostly smooth, or as in the case of the angular 
element of the mandible, marked only with a few comparatively coarse distant 
and prominent ridges. 

I take the liberty of dedicating this species to Professor Huxley, to whose 
researches ichthyological science is so much indebted for a more correct 
insight into the definition and structure of the Ccelacanthidaj. 

Suborder Acipenseroidei. 

Family Pal^eoniscid^. 
Genus Elonichthys, Giebel, 1848. 

(Giebel, Fauna der Vorwelt, vol. i. pt. 3, p. 249; Traquair, Carboniferous Ganoids, p. 47 ; 
and Quar. Journ. Geol. Soc. London, vol. xxxiii. 1877, p. 553.) 

Elonichthys serratus, sp. nov. 

PI. I. figs. 5-8. 

Two specimens only of this interesting form are contained in the Survey Col- 
lection, and both are unfortunately not quite perfect. The larger (fig. 6) measures 

* Agassiz, Poissons Fossiles, vol. ii. pt. 2, p. 173. 
t Huxley, Dec. Geol. Survey, vol. xii. p. 24. 



REPORT ON FOSSIL FISHES. 23 

3| inches in length, but the extremities both of the head and of the upper lobe 
of the caudal fin being deficient, the total cannot have been less than 4 
inches. The other (fig. 5), which wants the greater part of the head as well 
as of the caudal fin, represents a fish of slightly smaller dimensions, whose 
original length was probably 3^ inches. 

The shape is fusiform, moderately deep between the head and dorsal fin, 
thence tapering gracefully towards the tail. Some remains of the cranial roof 
bones in the parietal region show a closely granulated external surface, while 
the facial bones and those of the shoulder girdle are ornamented with wavy 
sub-parallel ridges. The scales are somewhat small, and, as usual, diminish in 
size and increase in obliquity towards the tail. On the flank scales (fig. 7) the 
ornament consists in the first place of very delicate closely placed grooves 
or furrows, often interrupted and intercalated, their direction in the upper 
part of the area being more or less oblique from above downwards and 
backwards, while in the lower they become parallel with the inferior margin. 
Towards the posterior margin a number of coarser foldings or elevations of the 
surface make their appearance, and presently end in about half a dozen rather 
strongly marked marginal denticulations. It may be mentioned that the 
minute striation is much less obvious on the smaller than on the larger of the 
two specimens. Posteriorly (fig. 8), as the scales become smaller, the ridging and 
striation become less prominent, and reduced to a few longitudinal grooves and 
punctures, which finally disappear near the commencement of the caudal fin, 
while at the same time the denticulations of the posterior margin become also 
fewer in number, and likewise ultimately disappear. The variations in the 
form of the scales on different parts of the body are in accordance with 
what is found in Elonichthys striolalus, Robisoni, &c. 

A considerable portion of the pectoral fin, the rays and their joints being 
however slightly dislocated, is preserved in the larger specimen, and affords 
sufficiently clear evidence that the principal rays of this fin were articulated up 
to their origins, the joints being rather longer than broad. A small ventral is 
shown in the other specimen. The dorsal fin is placed nearly opposite the 
interval between the ventrals and the anal; it is not of very large size, 
and is short-based, acuminate, high in front, and concavely excavated 
posteriorly; the anal, situated between the ventrals and the caudal, is 
similar in size and shape. Both of these fins have their rays divided by 
transverse articulations, for the most part rather distant, and their ganoid 
joints are marked by one or two longitudinal ridges and furrows, sometimes 
also a tendency to more minute striation is observable. Dichotomisation 
of the rays occurs towards their extremities. The caudal is deeply cleft ; 
the lower lobe of moderate size ; its rays, nearly quite smooth externally, 
are divided by distant articulations, which become, however, closer in the 

VOL. XXX. PART I. E 



24 RAMSAY H. TRAQTJATR'S 

fringing rays of the upper lobe. Minute fulcra may be seen wherever the 
anterior margin of a fin is perfectly preserved. 

Remarks. — The relative position, as well as the shape of the dorsal and 
anal fins, and the nature of the scale ornament, prove unmistakably that this 
little fish belongs to the group of species {Elonichthys Robisoni, striolatus, &c), 
which is so characteristic of the Lower Carboniferous rocks of Scotland. 
Although the hitherto described species of this group are in many cases difficult 
to distinguish, yet, in all, the serration of the posterior margin of the flank 
scales is minute, in some {Elonichthys tenuiserratus, Traq.) excessively so; 
here, the comparatively coarse and prominent aspect of these serrations forms a 
tangible mark of distinction ; the median fins are also proportionally smaller. 
So far as the present investigations go, the rare occurrence of representatives 
of the Robisoni group is certainly a remarkable feature in the palaeontology of 
the Eskdale beds. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 

Elonichthys pulcherrimus, sp. nov. Traquair. 
PI. I. figs. 9-12. 

Of this there is only one specimen in which the posterior part of the ventral 
margin is unfortunately cut off by a joint, so that the anal and caudal fins are 
deficient, although the commencement of the caudal body-prolongation is pre- 
served. The entire length exhibited is 4^ inches ; when perfect the fish could 
not have measured less than 5^ ; its greatest depth at the arch of the back, 
just in front of the dorsal fin, is 1^ inch. The body is thus rather deeply 
fusiform, and the length of the head (1| inch) is contained about five times in 
the estimated total. The cranial roof bones are covered with a small close 
tuberculation ; the facial bones exhibit a ridged ornamentation. The lower 
margin of the maxilla is, however, tuberculated, and is set with strong, sharp, 
conical teeth of different sizes, large and small ; one of the larger teeth measur- 
ing about -fa inch in length. The dentary element of the mandible is orna- 
mented with closely set ridges, mainly following the longitudinal direction of 
the bone, though also slightly divergent from back to front. The suspensorium 
is oblique and the gape wide, but the state of preservation of the specimen 
hardly allows of any further description of the bones of the head. 

The scales are of moderate size, and over the whole body highly ornate. 
Their exposed area is covered with close, delicate, yet sharply defined ridges, 
which mostly proceed from before backwards and end in fine denticulations of 
the posterior margin. On the flank scales (fig. 10) these ridges tend to proceed 
obliquely downwards and backwards on the upper and posterior part of the 
area, while below this, on the anteroinferior part, their direction is more 



REPORT ON FOSSIL FISHES. 25 

parallel with the inferior margin. Passing towards the caudal region (fig. 11), the 
ridges become less divisible into two sets, and are generally tolerably parallel 
with the upper and lower margins of the scale ; they exhibit, moreover, a 
tendency to reticulation or anastomosis, till finally, on the small lozenge-shaped 
scales of the caudal body-prolongation the pattern assumes more of a punc- 
tured than of a striated aspect. The denticulation of the hind margin of the 
body-scales is persistent up to the tail pedicle. A few strong, broad, imbricat- 
ing scales are seen in front of the dorsal fin, and the upper margin of the 
caudal body-prolongation is set with the usual median row of imbricating V 
scales, displaying a ridged ornamentation corresponding with that of the body 
scales. 

The dorsal fin commences midway between the head and the probable 
origin of the caudal, and is triangular and acuminate in shape, with the 
posterior margin excavated. Not less than thirty rays are contained in it, 
these being rather delicate, dichotomising towards their extremities, and 
divided by transverse articulations, which are somewhat distant. Externally 
the rays are ganoid, and finely striated in the direction of their length (fig. 12) ; 
the anterior margin of the fin is set with fine fulcra. None of the other fins 
are preserved. 

Remarks. — Elonichtliys pulcherrimus evidently belongs to the same group 
of species as Elonichtliys Egertoni of the British Coal Measures, and those 
originally described by Giebel {Elonichtliys Germari, &c), but it is so distinct 
in its short deep form and its beautiful scale ornamentation as to render 
detailed comparison unnecessary. 

Position and Locality. — From the Cement-stone group of the Calciferous 
Sandstone series, near Glencartholm, Eskdale. 

Genus Rhadinichthys, Traquair, 1877. 

(Traquair, Quar. Journ. Geo!. Soc. xxxiii. 1877, p. 558.) 

Rhadinichthys Geikiei, Traquair. 
Rhadinichthys Geikiei, Traquair, Proc. Roy. Soc. Edin. ix. 1877, p. 438. 

PL I. figs. 13-18. 

Description. — Length of a particularly fine and perfect specimen, 4^ inches ; 
less perfect examples, however, indicate that it often attained larger dimensions, 
and one in particular, judging from the size of the head, must have originally 
measured over 6 inches. The shape is elegantly fusiform and rather slender ; 
the length of the head being rather greater than the depth of the body at the 
ventral fins, and contained about 4-^ times in the total. 

The cranial roof bones are ornamented with closely set, well-defined 



•26 RAMSAY H. TEAQUAIR'S 

elevated ridges, which, although wavy and sometimes contorted, mainly follow 
a longitudinal direction, especially on the parietal and frontal plates and the 
posterior part of the superethmoidal ; they are often bifurcated, intercalated, and 
interrupted, and ou the lateral parts of the cranial shield these interruptions 
often become so frequent as to cause the ornament to assume a somewhat 
tubercular character ; in some specimens this condition invades even the more 
central parts. On the anterior part of the superethmoidal, which forms the 
usual projection over the mouth, the ridges are disposed transversely, or at 
right angles to those behind. 

The direction of the suspensorium is very oblique, and the gape conse- 
quently wide. The opercular bones are well developed, and of the usual shape 
in this genus, but their external markings are not distinctly preserved. The 
maxilla has its broad portion ornamented with fine ridges, which run parallel 
with its superior and posterior margins ; its infra-orbital process is tubercu- 
lated, and the tubercles are continued backwards as a narrow band along the 
entire inferior or dentary margin of the bone. The beautifully tapering- 
mandible is covered externally with fine ridges, which pass forward from the 
angle in a somewhat radiating manner, so that below they are pretty parallel 
with the inferior margin, while above they cut the superior one at low angles. 
In large specimens these ridges break up into a minute tuberculation along the 
superior margin of the jaw — a condition rarely met with in the smaller 
examples. The orbit is anteriorly placed, and furnished with the usual arrange- 
ment of suborbital and circumorbital plates. 

The bones of the shoulder girdle are ornamented with well-defined sub- 
parallel ridges, which are arranged according to the common pattern, namely, 
somewhat concentrically, and more or less parallel with the margins of the 
bones. 

The scales are of moderate size, with narrow covered area, rhomboidal in 
shape, and increasing in obliquity, and diminishing in size towards the tail. 
On the flanks they are tolerably equilateral, but along the belly between the 
pectoral fins and the posterior part of the base of the anal, they suddenly 
become very low and narrow, their height diminishing to from ^ to ^ of their 
length. This change of form usually occurs at the fourth or fifth scale down- 
wards from the lateral line. A few large scales are seen in front of the dorsal 
anal, and lower lobe of the caudal fin, and on the body-prolongation in the 
upper lobe of the latter the usual modifications of shape occur. 

As regards the scale-markings, their general plan may be described as 
follows : — Taking a scale from the anterior part of the flank (figs. 15, 16), a few 
delicate grooves are seen passing down along the anterior margin of the ganoid 
area, which then turn round at the anterior inferior angle so as to become 
parallel with the inferior margin. The rest of the area presents, according to 



REPORT OF FOSSIL FISHES. 27 

the size of the specimen, from 4 to 7 longitudinal elevations or feeble ridges 
running across the scale with a slight downward obliquity, as well as a 
slight tendency to radiation, and ending on as many sharp denticulations of 
the hinder margin. Towards the back and belly (fig. 18) both sets of markings 
tend to become confounded into one set of delicate, more or less diagonal ridges 
and grooves, and towards the tail (fig. 17) these markings usually fade away ; 
a few punctures and longitudinal furrows being in most cases all that is to be 
seen on the scales behind the dorsal and anal fins. In different individuals, 
however, very considerable differences are found as regards the strength and 
prominence of the scale-markings. In some the markings are very distinctly 
and sharply defined (fig. 15), and the scales continue to be highly ornate up 
to the caudal body-prolongation, while in others (fig. 16) even those on the 
front part of the body are comparatively smooth, the middle of the scale 
being marked with a few short grooves, and the oblique ridges becoming- 
apparent only towards the posterior margin ; a reference to the figures 
will, however, give a better idea of the variations in the general aspect of 
the scale sculpture than any amount of description. 

The pectoral fin is rather small, its length being hardly more than half 
that of the head ; the larger rays on the praeaxial aspect are unarticulated 
till towards their terminations. The ventrals are small and delicate, and not 
well exhibited in any of the specimens. The dorsal fin commences only very 
slightly in front of the anal, the two fins being thus placed nearly opposite 
each other ; they are very similar in shape and structure, being moderate in 
size, acuminate, concavely cut out behind, and composed of delicate rays, 
which dichotomise towards their extremities, are distantly articulated, and 
having their brilliantly ganoid joints marked usually by a single sharp 
longitudinal furrow, though even this is sometimes wanting save near the 
origin of the fin. Between the anal fin and the commencement of the lower 
lobe of the caudal is an interval fully equal to the length of the base of the 
former. The caudal is also moderate in size, deeply bifurcated, and with a well- 
developed body-prolongation along the upper lobe ; its rays partake of the same 
general appearance as those of the dorsal and anal, though their articulations 
are a little closer, and their surfaces usually quite smooth. Delicate fulcra 
are observable in the anterior margins of all the fins. 

Var. elegantulus (PI. II. figs. 1-5). — Length 2\ to 2| inches, elegantly fusi- 
form, general proportions as in the foregoing, but the cranial roof bones are 
ornamented with closely set rounded ridges, which are proportionately some- 
what coarser, and more contorted in their arrangement ; in one very perfect 
specimen, in which the head bones are beautifully exhibited, the maxilla is 
destitute of tuberculation along its inferior margin. The markings on the scale 
are faint, those on the sides and belly being nearly smooth ; their posterior 



28 RAMSAY H. TRAQTJAIR'S 

margins display a few denticulations — 4 to 5 on the largest flank scales, 
diminishing to 2 or 3 in those further back. The most remarkable feature in 
this form is, however, the very small development of the low narrow ventral 
scales, which are so conspicuous in the preceding. Such scales are indeed 
distinguishable along the ventral margin between the pectoral and anal fins, 
but on the other hand, in the abdominal region as many as eight scales may 
be counted down from the lateral line without any prominent change in form 
taking place. 

Remarks!. — This is by far the most common fish in the Eskdale beds, 
and from the comparatively large number of specimens before me, has 
admitted of a very full description, nevertheless there are difficulties in the 
way of its satisfactory determination as a species. 

In 1877, I described, under the name of Rhadinichihys Geikiei, a small 
fish from the Wardie Shales near Colinton, Mid-Lothian, from specimens, 
which unfortunately were mostly fragmentary or distorted, save one very small 
one, which I considered to be a young individual of the species. These 
specimens, so far as they go, show a fish of much the same proportions as 
the above described form from Eskdale ; the markings on the head bones, 
where visible, are very similar ; so is the scale ornament, though perhaps 
the ridges are a little finer, and the denticulations of the posterior margin 
rather more numerous in proportion to the size of the scale. None of the 
larger specimens show the ventral region, save one, in which a similar 
arrangement of low narrow scales exists, though not so conspicuously, and 
the absence of this feature in the smaller examples is paralleled by the condition 
found in those from Eskdale, which I have felt obliged to consider only as 
young forms, or at most as a variety {elegantulus). On the whole, after 
most careful comparison of every scrap from both localities, I cannot find any 
very tangible or decisive mark of specific distinction, and therefore, although I 
may possibly be hereafter proved to be in error, I do not feel justified in 
separating the common Rhadinichthys of Eskdale from Rhadinichthys Geikiei. 

Again, it is possible to point out scales on the flanks of many Eskdale 
specimens of Rhadinichthys Geikiei which are indistinguishable from those of 
the Coal Measure Rhadinichthys monensis, Egerton. This isolated fact, 
however, cannot prove the identity of the two species unless corroborated by 
the discovery of more perfect examples of Rhadinichthys monensis, which is 
as yet only known from scattered scales, and very fragmentary specimens 
indeed. What I have seen of the latter certainly inclines me to believe in the 
distinctness of the two forms, and the similarity of certain scales is of constant 
occurrence in closely allied species of Palpeoniscida2. 

Position and Locality.- Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 



JREPORT ON" FOSSIL FISHES. 29 



Rhadinichthys delicatulus, sp. nov. 
PI. II. figs. 6-9. 

Description. — Length from 1^ to 3^ inches ; shape and general proportions 
as in the last described species. 

The bones of the head and shoulder girdle are externally sculptured with 
delicate ridges, which, although they follow the same general pattern as in 
Rhadinichthys Geikiei, show less tendency to contortion and interruption, and, 
except along the dentary margin of the maxilla, are nowhere seen to break up 
into tubercles. The scales (figs 8, 9) are proportionately thinner, and those of 
the front part of the body have their upper and lower margins rather straighter 
and more parallel with the long axis of the body, but the same arrangement 
of low narrow scales is seen along the belly. The vertical grooves along the 
anterior margin of the sculptured area of the scales are almost lost, nearly the 
whole surface being covered with minute sharp ridges and furrows, which, with 
the exception of one or two which run parallel with, and close to, the 
inferior margin, are directed rather diagonally across the scale from before 
backwards, ending on fine denticulations of the hinder border ; sometimes two 
of these ridges ending on one clenticulation. Finally, this delicate ornament 
is in most cases continued back to the scales of the tail pedicle itself. The 
fins are similar in position, shape, and structure to those of Rhadinichthys 
Geikiei, but, comparing specimens of the same size, their rays seem to be 
rather more delicate, and more distantly articulated. One specimen (fig 6), the 
largest of the series, shows the strange phenomenon of the upper lobe of 
the caudal fin being neatly cut off and laid across the lower one. 

Remarks. — Rhadinichthys delicatulus so closely resembles the preceding 
species in structure and proportions, that I was for long in great doubt as to 
whether it were not better to treat it as a mere variety, or perhaps, seeing that 
the specimens are mostly of small size, as a young form. But so far as the 
collection goes, the differences between the scale-markings of the two forms 
are so constant that it is always easy to point out the specimens referable to 
the one and to the other. Moreover, small specimens, both from Colinton and 
Eskdale, which I believe to be referable to the young of Rhadinichthys Geikiei, 
have the scales always comparatively smooth, whereas in Rhadinichthys deli- 
catulus, the smaller the specimen, the more decided appears the distinctive 
pattern of the scale-markings. On these grounds I have decided to consider 
Rhadinichthys delicatulus as a " good species," in the sense in which that 
term is usually employed. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 



30 RAMSAY H. TRAQUAIR'S 

Rhadinichthys Macconochii, sp. no v. Traquair. 
PI. II. figs. 12-16. 

Description. — Length from 3 to 3^ inches ; shape elegantly fusiform, the 
dorsal and ventral lines being gently and evenly curved. The length of the 
head is contained a little more than four times on the total, and is equal to the 
depth of the body midway between the pectoral and ventral fins. 

The cranial roof bones are ornamented with a close, comparatively coarse, 
and frequently confluent tuberculation ; the orbit is, as usual, anteriorly placed, 
and the ethmoid forms a projection over the mouth. The suspensorium is 
very oblique, and the gape correspondingly extensive. The maxilla is of the 
usual form, its broad portion being ornamented with closely set ridges, which 
run parallel with its superior and posterior margins ; the beautifully tapering 
mandible is marked externally with ridges which pass from behind forwards in 
a slightly radiating manner, but which are also so frequently interrupted as to 
cause the ornament to assume nearly as much of a tuberculated as of a striated 
aspect. The operculum is of moderate size, rather broader below than above ; 
the interoperculum is rather large ; both of these plates are ornamented with 
prominent and proportionally coarse rugae, which run mostly parallel with the 
lines of growth. 

The bones of the shoulder girdle present nothing peculiar in form and 
arrangement, and are sculptured externally with ridges similar to those on the 
opercular bones. 

The scales are of medium size, rhomboidal, as usual diminishing in size 
dorsally, ventrally, and posteriorly ; they are low and narrow on the belly from 
the throat to the anal fin, while those of the front part of the lateral line are 
proportionally higher than the others. The scales of the middle line of the 
back are small until just in front of the dorsal fin, where a few of comparatively 
large size and imbricating arrangement are found. In one specimen 45 
oblique dorso-ventral bands of scales may be counted from the shoulder 
girdle to the commencement of the lower lobe of the caudal fin. The scale 
ornament is sharpest on the scales above the lateral line, where it consists first 
of a few sharp grooves parallel with the anterior margin, and tending below to 
turn round along the inferior one, the rest of the area being occupied by two or 
three slightly prominent ridges, passing somewhat obliquely across towards 
the posterior margin, before reaching which they usually stop short, a pro- 
minent feature in this species being that on no part of the body do the scales 
appear to be denticulated posteriorly. Towards the tail the vertical furrows 
become imperceptible, or reduced to a single one. Below the lateral line the 
scale ornament is for the most part less marked, though similar in character, 
but towards the tail pedicle little or no difference is seen above and below. 



REPORT ON FOSSIL FISHES. 31 

The "pectoral fin is scarcely more than half the length of the head ; its 
principal rays are unarticnlated till towards their terminations. The situation 
of the ventrals is indicated by a few stumps of rays at a point a little behind 
the middle point between the pectorals and the anal. The dorsal is situated 
far back, being nearly exactly opposite the anal ; both of these fins are similar 
in appearance, being moderate in size, triangular, acuminate, and slightly cut 
out behind ; their rays are of medium coarseness, smooth, dichotomising 
towards their extremities, and having their articulations somewhat distant. 
The caudal is well developed, heterocercal, deeply cleft, the rays similar in 
appearance to those of the dorsal and anal ; in the lower lobe dichotomising 
towards their extremities, in the upper towards the middle. Traces of well 
developed fulcra are seen along the fin margins. 

Remarks. — The position of the above described species in the genus 
Rliadinichthys is indicated by the structure of the pectoral fin, by the shape of 
the scales and the nature of their sculpture, and by its general form and pro- 
portions, although the dorsal fin is placed still further back than is usual in 
Rliadinichthys, its position being hardly if at all in advance of the anal. This 
latter character allies it to Cyclopty chins, but the peculiar form of scale with 
the posterior inferior angle rounded off, and which constitutes one of the main 
diagnostic marks of the last named genus, is here absent. Its main specific 
characters — the tuberculation of the cranial shield, the peculiar sculpture of 
the mandible, and the non-denticulation of the posterior margins of the scales 
taken along with its size and proportions — are so exceedingly well marked, 
that it may at once be identified even from small fragments. 

I have pleasure in naming this new species after Mr Arthur Macconochie, 
Fossil Collector to the Scottish Geological Survey, to whose industry in his 
department is due the discovery of the rich deposits of new fishes and Crustacea 
in the Lower Carboniferous rocks of Eskdale and Liddisclale. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 



Rliadinichthys tuber cidatus, sp. nov. Traquair. 

PI. IV. figs. 1-3. 

Description. — The length of the only entire specimen which has occurred, 
is 7 inches from the tip of the snout to the bifurcation of the caudal fin ; the 
extremity of the upper lobe of the tail is not preserved, otherwise the total 
length of the fish would probably be, at least, 8-| inches. The length of the 
head is 2 inches, equalling the greatest depth of the body just in front of the 
ventral fins, and being contained little more than three times in the length of the 

VOL. XXX. PART I. F 



32 RAMSAY H. TRAQUAIR'S 

specimen up to the bifurcation of the caudal, or more than four times in the 
estimated total. The depth of the tail pedicle is f inch. 

The head is much crushed, and its bones badly preserved, its structure is, 
however, clearly seen to be typically palreoniscoid, with very oblique suspen- 
sorium, anteriorly placed orbit, wide gape, and powerful jaws. The operculum 
seems somewhat long and narrow, the interoperculum square-shaped. No 
teeth are visible. 

The bones of the head being almost everywhere seen only from their 
internal surfaces, their external ornamentation is but scantily exhibited. 
Evidences of a minutely tubercular sculpture, the tubercles being sometimes 
rounded, sometimes slightly elongated or confluent, are seen on the parietal 
and ethmoidal regions of the cranial roof as well as, in impression, on a small 
portion of the interoperculum. Towards the extremity of the mandible also, 
a patch of the outer surface of the bone is seen, but here the ornamentation 
consists of closely set delicate wavy ridges running in a longitudinal direction. 

The bones of the shoulder girdle — supra-clavicular, clavicle, and infra- 
clavicular — are pretty well shown, the two latter from their internal aspects 
only. The outer surface of the supra-clavicular displays some traces of a 
longitudinal striated sculpture. 

The scales are rather small for the size of the fish, especially at the tail, to- 
wards which region they rapidly diminish. In the flank scales (fig. 2) the covered 
area is narrow ; the sculptured one presents a few fine ridges and grooves 
along the anterior margin, the rest of the space being covered with small 
closely set tubercles, sometimes rounded, sometimes elongated or confluent. 
Towards the tail (fig 3), and also towards the dorsal and ventrical margins, the 
tuberculations largely gives way to a ridged ornamentation; the ridges running 
parallel with the anterior and inferior margins, sometimes also with the 
superior, while the postero-central portion of the area is occupied by tubercles, 
tending to become confluent, with diagonal ridges which are a little coarser, 
and more wavy and irregular than the marginal ones already mentioned. 
Powerful longitudinally ridged V-scales protect the upper margin of the caudal 
body-prolongation, which is of great strength, and the acute lozenge-shaped 
scales which clothe its sides are of small size, arranged in many rows, and, so 
far as they are traceable, ornamented with sharp diagonal ridges. 

On the attached surface of the body-scales, the vertical keel is rather 
delicate, yet very distinctly defined, as is also the socket for the reception of 
the articular spine of the scale next below ; this spine, which is of moderate 
size, arising from the upper margin of the scale close behind the upper termina- 
tion of the keel. As usual, spines and sockets disappear in the scales of the 
posterior part of the body. No denticiilations are observable on the posterior 
margins of any of the scales. 



REPORT ON FOSSIL FISHES. 33 

The pectoral fin is 1| inch in length, and has its principal rays unarticulated 
for the greater part of their length. The origin of the ventral is situated 1 T 5 ^- 
inch behind that of the pectoral ; the fin itself is not in good preservation, its 
rays being much broken up, so that its shape and size are lost. The dorsal fin 
is situated nearly opposite the anal, arising opposite a point 4 inches back from 
the tip of the snout ; this arises only a little behind the dorsal, and extends 
correspondingly farther back. Both of these fins are rather large, each 
measuring about an inch in length at the base and the same in height 
anteriorly, and the latter measurement would probably be somewhat greater 
were their larger rays preserved up to their extremities, which does not seem 
to be the case. They are also very similar in shape, being triangular and 
acuminate ; their numerous and rather delicate rays have their transverse 
articulations somewhat distant, so that the joints are rather longer than broad ; 
no sculpture is visible. The caudal is incompletely preserved, the extremities 
of both lobes being wanting, but enough is seen to show that it was powerfully 
developed, deeply cleft, and having a body-prolongation of great strength along 
the upper lobe ; the rays are similar in character to those of the dorsal and 
anal. Well developed fulcra are seen in connection with the margins of the 
fins wherever these margins are preserved. 

Remarks. — I have considered this strikingly new and beautiful Palaeoniscid 
to be a Rhaclinichthys on account of the structure of the pectoral, and the 
relative positions of the dorsal and anal fins, although the caudal body- 
prolongation is more powerfully developed than in the more typical members of 
the genus, such as Rhadinichthys ornatissimus, Rhadinichthys carinatus, &c, and 
although the epithet " slender fish " can hardly be applied to its proportions. 
Its large head, short, thick, fusiform body, peculiar ridged-tuberculate scale 
ornament, and non-denticulated scales, with other peculiarities which need not 
be recapitulated, form an assemblage of specific characters which collectively 
are so novel that no detailed comparison with any other species is necessary. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 

Rhadinichthys (?) angnstuhis, sp. nov. Traquair. 

PI. II. figs. 10, 11. 

Two specimens only of this interesting and somewhat doubtful form have 
occurred, one of which, the larger and less perfect, measures 2^ inches in 
length, while the other more perfect example attains a length of only 1^ inch. 
The length of the head is equal to about £ of the total ; the greatest depth of 
the body is at the shoulder, and is contained about six times in the entire 
length of the fish, while it is not so much as twice the depth of the tail pedicle, 
the dorsal and ventral margins being nearly straight. The general contour is 



34 RAMSAY H. TRAQTJAIR'S 

therefore peculiarly short and straight, and wanting in the usual more or less 
fusiform outline, while the tail pedicle is of great proportional depth. 

All that can be said of the head is that it is typically palseoniscoid in 
structure, with oblique suspensorium, &c, and that some traces of a minute 
ridged ornament are seen on some of its delicate bones, e.g., the mandible. The 
body-scales (fig. 1 1) are of moderate dimensions in proportion to the size of the 
fish, and are marked each with three or four delicate, yet sharply-defined and 
somewhat distant ridges, which run right across from before backwards, parallel 
with the superior and inferior margins ; on the minute lozenge-shaped scales of 
the caudal body-prolongation these ridges, now excessively fine, are diagonal in 
position ; the V-scales of the tail are proportionally largely developed. 

The pectoral and ventral fins are small, the dorsal and anal nearly opposite, 
though the former arises a little in advance of the latter. The two last named 
fins resemble each other in their triangular-acuminate contour ; the caudal is 
not completely preserved, but its appearance seems to indicate that it was 
bifurcated in the usual manner. The fins are preserved only in the smaller of 
the two specimens, and their rays are so excessively delicate that it is im- 
possible to describe their articulations, but they are closely set, and appear to 
bifurcate towards their extremities. 

Remarks. — On account of its general structure, so far as can be made out, 
along with the form and position of the fins, this strange little fish is referable, at 
least provisionally, to the genus Rhadinichthys. Its prominent specific characters 
are its large head, short straight body, deep tail pedicle, and the scale sculpture 
consisting of a few delicate, straight, non-bifurcating longitudinal ridges. The 
scale ornament of Rhadinichthys Grossarti, Traq.,* another very small species 
from the Coal Measures of Lanarkshire, also consists of straight ridges, but 
these are more or less oblique in their direction, besides being closer and more 
numerous, while the shape of the fish, with its narrow elongated tail pedicle, 
forms a character at once distinguishing it from the form before us. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cementstone 
group of the Calciferous Sandstone series. 

Rhadinichthys (?) fusiformis, sp. nov. Traquair. 

PI. III. figs. 1-5. 

Description. — Length of an entire specimen, 6^ inches ; shape elegantly 
fusiform ; length of the head equal to the greatest depth of the body between 
the shoulder and dorsal fin, and contained 4| times in the total, 3^ times up 
to the bifurcation of the caudal fin. The dorsal fin is placed far back, so as to 
be nearly opposite the anal ; the caudal is very heterocercal and inequilobate, 

* Proc. Roy. Phys. Soc. Edin. vol. iv. pt. 3, 1878, pp. 237-245. 



REPORT ON FOSSIL FISHES 35 

the length of the upper lobe, from a point opposite the commencement of the 
rays of the lower one, being 2f inches. These proportions are taken from a 
very perfect and undistorted specimen, whose entire length is 6^ inches ; none 
of the others are quite perfect or free from distortion, though the peculiar 
characters of the fish are such as to enable it to be easily recognised even 
from small fragments. 

The head is typically paheoniscoid in structure, with oblique suspensorium, 
wide gape, ethmoidal prominence, and anteriorly placed orbit. The cranial 
roof bones are ornamented with closely set irregularly contorted rugse, 
frequently interrupted, so as at times to pass into tubercles. The operculum 
is narrow and oblong, the interoperculum, as usual, short and quadrate ; these 
plates are in all cases ill preserved, so that little can be said of their external 
sculpture, save that it seems to be of a striated character. The maxilla is 
of the usual paheoniscoid shape, and has its broad postorbital portion covered 
with wavy and contorted ridges, which in most instances pass into a narrow 
band of irregularly shaped tubercles stretching along the dentary margin. 
The mandible is very stout, its depth posteriorly equalling f of its length, in 
shape it rapidly tapers towards the symphysis. Externally it is covered with 
closely set, slightly wavy ridges, which, running from behind forwards, diverge 
from each other along a longitudinal line placed rather below the middle of the 
bone, on whose upper and lower margins they obliquely impinge, but the 
strice forming the lower side of this somewhat feather-like pattern are much 
more horizontal in direction than those on the upper. The jaws are armed 
with conical teeth of two sizes, large ones being placed at short intervals 
inside a row of minute external teeth. 

The bones of the shoulder girdle are striated with tolerably coarse wavy 
ridges, which on the upper or vertical part of the clavicle are again fretted with 
minute transverse indentations. 

The scales of the body are of moderate size, rhomboidal, and tolerably thick. 
On the front part of the flank (figs. 2,3) they are tolerably equilateral, with slightly 
concave upper and convex lower margin ; the covered area is very narrow, the 
articular spine moderate in size, and the keel of the attached surface only 
slightly developed. Towards the tail (fig. 5) and along the back the scales 
become smaller and more oblique, and in front of the dorsal fin there are four 
or five imbricating median scales of a larger size. Along the belly (fig. 4) they 
become very low and narrow, and on the caudal body-prolongation, as usual, 
small and acutely lozenge-shaped, while imbricating V-scales clothe the upper 
margin of this part. The scales are marked externally by a very ornate and 
easily recognised sculpture, though it is excessively difficult by means of words 
to give anything like an adequate description of its peculiarities. It consists 
of sharp furrows or grooves, sometimes interrupted and intercalated, some of 



36 RAMSAY H. TBAQUAIR'S 

which run parallel with the anterior and inferior margins, while others run 
more or less diagonally across the remaining portion of the sculptured area. 
According to the flatness or elevation of the interspaces between these furrows, 
a greater or less appearance of ridging is produced in different specimens, and 
in all the ridged appearance is pretty strongly developed in the scales of the 
back between the dorsal fin and the occiput. The ornament becomes less 
sharp posteriorly, but nevertheless it is developed to a greater or less extent 
even on the scales of the caudal body-prolongation. Some amount of a 
tolerably coarse denticulation is also observable, especially on the flank-scales, 
and, as very commonly happens, disappearing towards the tail and the margin 
of the body. 

The pectoral fin is tolerably preserved only in one specimen, and is 
acuminate in shape ; its length equals about f that of the head. So far as can 
be made out by careful examination with a good lens, its principal rays seem 
to be articulated up to very near their origins, — a feature which is certainly 
at variance with the characters of the genus Rhadinichthys, hence the query 
which I have appended above. The ventral fin, situated between the pectoral 
and anal, is ill preserved ; it seems however to have been small, with its rays 
moderately closely articulated, and fretted with a striated pattern. The dorsal 
fin is placed behind the arch of the back, its anterior commencement being 
only very slightly in front of that of the nearly opposed anal ; it is moderate in 
size, acutely triangular-acuminate in shape, the extent of its base measuring 
only about half that of its height in front. Its rays are about 30 in number, 
delicate, bifurcating towards their extremities, their joints longer than broad, 
but becoming shorter distally, and in the hindermost rays ; they are externally 
smooth, or with a single longitudinal furrow. The anal is not so well preserved 
in any of the specimens, but is apparently of the same form and structure as the 
dorsal. The caudal is powerful, deeply cleft, and inequilobate, the upper lobe 
being nearly twice as long as the lower ; its rays are delicate, smooth, 
dichotomising towards their extremities, and divided by tolerably distant 
transverse articulations. The anterior margins of all the fins are minutely 
fulcrated. 

Remarks. — The characteristic features of the above described beautiful 
Palseoniscid are so distinct and so novel, that we are fortunately free from any 
troublesome doubts and questions as to species, for even although the scale 
markings may show some amount of individual variation as to the strength and 
sharpness, it is, as already stated, easy to pick out its fragmentary remains and 
place them together as belonging to one well-defined form. Superficially it 
reminds us of the well known and typical Rhadinichthys ornatissimus of the 
Lower Carboniferous rocks of Central Scotland, but the differences in scale-sculp- 
ture and general proportions come into such strong relief the moment a critical 



REPORT ON" FOSSIL PISHES. 37 

examination is entered into that a detailed comparison is quite unnecessary. 
Indeed, the structure of the pectoral fin seems to forbid its being placed in the 
genus Rhadinichthys at all, for that member, so far as can be seen, has its rays 
articulated as in the species of Elonichthys of the " Robisoni " type, from which 
it is however excluded by the backward position of the dorsal fin. However, 
rather than hastily to proceed to the creation of a new genus upon that 
character alone, I have placed it provisionally in Rhadinichthys, which it 
certainly resembles more than any other in general contour. 

Genus Cydoptychius (Huxley), Young, 1865. 

(Young, British Assoc. Reports, 1865, vol. xxxv. p. 318; Traquair, Geol. Mag. Decade II. 
vol. i. No. 6, June 1874). 

Cycloptychius concent ricus, sp. no v., Traquair. 
PL II. figs. 17-20. 

Description. — The largest specimen attains a length of 4| inches, if we allow 
for a small portion of the front of the head which is broken off. The length of 
the head is contained about 5 times, the greatest depth of the body 6^ times, 
in the estimated total, up to the extreme end of the upper caudal lobe. The 
contour of the fish is therefore peculiarly slender and graceful, the depth of the 
body continuing pretty uniform as far as the origin of the posteriorly placed 
dorsal fin, whence it tapers to the moderately stout tail pedicle. 

The head is somewhat elongated, with very oblique suspensorium, extensive 
gape, anteriorly placed orbit, and well marked ethmoidal prominence. The 
sculpture of the cranial roof bones is not exhibited ; on the other bones of the 
head it appears to be of a striated character, but it is only distinctly seen in 
the case of the maxilla and mandible. The former has its broad post-orbital 
portion covered with closely set ridges, which pass into an irregular tubercula- 
tion along the dentary margin ; the mandible, slender and tapering in shape, 
has also a narrow band of tuberculation along its upper margin, but below this the 
surface is striated with fine ridges, which proceed diagonally from above down- 
wards and forwards, and increase in obliquity from behind forwards. Small 
conical teeth of different sizes may be observed in several specimens. 

The bones of the shoulder girdle present nothing peculiar in their configur- 
ation, and, so far as their external sculpture is visible, it consists of wavy sub- 
parallel ridges. 

The scales of the flank (fig 19) are somewhat large for the size of the fish, and 
are higher than broad, with their posterior inferior angles obtuse; towards the back, 
belly, and tail, they become smaller, and assume the usual rhomboidal shape. 
The articular spine is well marked though small; the keel of the attached 



38 RAMSAY H. TRAQUAIR'S 

surface is central and sharply defined. The scale markings are peculiar and 
characteristic. Along the sides of the body, as far as the tail pedicle, the 
exposed surface of the scale is ornamented by somewhat coarse and closely 
placed ridges, which, commencing at the upper margin, descend in such a way 
as to form one median ridge, surrounded by several others, which running 
parallel with each other, and with the anterior and posterior margins, join 
each other below round the lower extremity of the median one ; or— to put the 
matter in another way — we have a set of concentric ridges parallel with the 
anterior, inferior, and posterior margins, with an odd one in the middle, or 
sometimes with two uniting in a loop ; in addition, there are often one or two 
fine vertical stride along the anterior margin. A somewhat different pattern is 
observable along the back (fig 18), extending downwards to, and including the 
second longitudinal row of scales above the lateral line. Here there is ordinarily 
only one marginal ridge, running closely along the anterior, inferior, and posterior 
margins, within which the area is, according to the size of the scale, occupied 
by from two to five diagonal and parallel ridges, passing from before down- 
wards and backwards. The caudal body prolongation is comparatively weak 
and narrow ; its minute lozenge-shaped scales are ornamented by diagonal 
ridges only ; stride of a similar description characterising also the large V-shaped 
scales which run along its upper margin. 

The pectoral fin is small, its length being hardly more than half that of the 
head ; it is acuminate in shape, and consists of about twelve rays,, of which the 
principal ones are unarticulated till towards their terminations. The ventrals 
are in no specimen well preserved, but seem to have been likewise small, and 
with their rays somewhat distantly articulated. The median fins are, on the 
contrary, of tolerably large size. The dorsal is situated far back, and is 
triangular-acuminate in shape ; its rays, the number of which cannot be accu- 
rately ascertained, aie slender, smooth, and distantly articulated. The anal 
may be said to be placed opposite the dorsal, though in some specimens it 
seems to commence slightly in front of it, and to be also somewhat larger ; it 
is sharply acuminate in front, with concavely excavated posterior margin ; the 
rays are of the same character as those of the dorsal. The caudal is well 
developed, deeply cleft, and inequilobate ; its rays resemble those of the other 
median fins. 

Remarks. — This is one of the most beautiful of the many beautiful fishes 
which the Eskdale strata have yielded, and nothing can be more strikingly new 
than its specific characters, of which the first which arrests the attention is 
the unusually bold sculpture of the scales, together with the peculiar form of 
those on the anterior part of the flank. Of hitherto described species, the only 
one which bears any resemblance to it is the Cycloptychius carbonarius of 
Huxley, the points of likeness being — the slender elongated form of the body, 



REPORT ON FOSSIL FISHES. 39 

the position and structure of the fins, the obtuseness of the posterior-inferior 
angles of the scales, and, last but not least, the possession of a peculiar scale 
ornament, consisting of ridges running parallel with the anterior, inferior, and 
posterior margins of the scale. These considerations indicate that Cydoptychius 
is the most appropriate genus to which to refer the present species, while, at 
the same time, the distinctions between it and Cydoptychius carbonarius are 
apparent at the first glance. In Cydoptychius carbonarius, the ridges on the 
scales are very much finer, that along the posterior margin being also more or 
less zigzagged in contour, while there is not that difference in the sculpture 
scales along the back which is so prominent a character in Cydoptychius con- 
centricus. The shape of the flank-scales also differs to a considerable extent ; 
for while in the former species the posterior-inferior angles are simply rounded 
off, in Cydoptychius concentricus they are absolutely obtuse, so that the inferior 
margin looks as much backwards as downwards. 

The genus Cydoptychius has not hitherto occurred below the horizon of 
the Millstone Grit. 

Position and Locality. — Near Glencartholm, Eskdale, in the Cement-stone 
group of the Calciferous Sandstone series. 

Phaner osteon, gen. no v. Traquair. 

Shape fusiform ; head typically palseoniscoid in structure ; body for the 
most part destitute of scales, so that the internal skeleton is well exhibited. 
Fins well developed ; anal commencing opposite the middle of the dorsal ; 
caudal heterocercal, only feebly bifurcated. 

Phanerosteon mirabile, sp. nov. Traquair. 

PL III. figs. 6-8. 

The entire length of the most perfect specimen is 2f inches, and in this 
measurement the length of the head is contained 4^ times. The cranial roof 
bones are granulated with minute and occasionally confluent tubercles. The 
suspensorium is oblique, the operculum oblong, with rounded-off posterior- 
superior angle, and showing traces of fine ridges, corresponding with the lines 
of growth ; the interoperculum is somewhat quadrate-rhomboidal, with convex 
posterior margin. The maxilla, of the typical palseoniscoid shape, has its post- 
orbital portion marked with fine ridges, running parallel with the posterior and 
superior margins ; the dentary margin and infra-orbital process appear to be 
finely tuberculated. The slender mandible displays on its outer surface 
numerous delicate ridges obliquely inpinging on its upper margin ; a few 
delicately striated branchiostegal plates are seen below it ; no teeth can be 

VOL. XXX. PART I. G 



40 RAMSAY H. TRAQUAIR'S 

distinctly made out on either jaw. The orbit is large, but is, as usual, 
anteriorly placed, and the ethmoid forms a rounded prominence above the 
mouth. The bones of the shoulder girdle present nothing peculiar in their 
configuration and arrangement. 

Immediately behind the upper limb of the clavicle, a few scales are visible, 
apparently the remnants of two or three dorso-ventral bands ; they are very 
indistinctly preserved, yet their shape appears to be rhomboidal, and somewhat 
higher than long ; and, though traces of ganoine appear on their surfaces, no 
sculptured pattern is visible. From this part, as far as the caudal fin, not the 
slightest trace of scales of any kind can be perceived,"" except in one specimen, 
where there are four well-preserved median scales in front of the dorsal fin. 
But in all, the caudal body-prolongation is furnished with scales, which are as 
solid and as well preserved as in any other ganoid of similar size from the same 
beds. Along the upper margin of this part we find the usual median row of 
pointed imbricating scales, and simultaneously with these there commences a 
band of lateral ones, clothing the side of the prolonged body axis, these being- 
very minute, acutely lozenge-shaped, and marked each with a few fine diagonal 
grooves ; this band of lateral scales does not, however, extend to the origin of 
the caudal fin rays until the base of the lower lobe is passed. A few imbri- 
cating median scales may also be seen just in front of the commencement of the 
lower lobes of the caudal fin. 

The absence of body scales reveals the internal skeleton in a manner 
unusually distinct for fishes of this family. There is no trace of vertebral 
bodies, the position of the persistent notochord being indicated by an empty 
space. Above this there is a series of short neural spines, bifurcated 
proximally, and slightly dilated distally, sixteen of which very regularly 
placed may, in one specimen, be counted between the head and the dorsal 
fin, beyond which they are a little confused ; but they are again seen in 
more undisturbed succession towards the tail, where they assume a much 
more backward inclination than in front. Again, above these the dorsal 
fin is supported by two sets of slender interspinous bones, proximal and distal. 
The proximal set are directly superimposed on the extremities of the neural 
spines, but they are more numerous and consequently more closely placed ; 
their exact number is not ascertainable, though I count thirteen of them to seven 
spines. Their distal extremities articulate with the proximal ends of the 
second set, with which they correspond in number ; the latter are somewhat 
shorter, and have both extremities somewhat dilated. 

On the haemal aspect of the notochordal space, there may be seen between 

* In the specimen figured, a tolerably large scale lies irregularly across the middle of the body ; 
but this, being evidently a dorsal ridge scale, is clearly out of its place, and probably belonged to some 
other fish. 



REPORT ON FOSSIL FISHES. 41 

the head and the dorsal fin, a series of ossifications whose exact form is not 
easy determinable, though some of them look somewhat V-shaped, and they 
may have served to enclose the aortic trunk. No trace of anything like ribs is 
observable. A little behind the commencement of the dorsal fin distinct 
haemal arches and spines appear as well as interspinous bones, supporting the 
anal ; but they are unfortunately, in the region of the last named fin, somewhat 
confused and mixed up, so that it is impossible to ascertain if its supporting 
ossicles were in double series, like those of the dorsal. Behind the anal, 
however, the hsemal spines are regularly disposed, and when they are seen 
supporting the lower lobe of the caudal fin, they are laterally flatted and dilated 
at their extremities ; further on they are concealed from view by the scales of 
the caudal body-prolongation. 

At the origin of each ventral fin, something like a small triangular pelvic 
bone is observable. 

The pectoral fins consist of very delicate rays, so delicate that, even with a 
powerful lens, it is difficult to decide as to the extent to which they are articu- 
lated. The small ventrals are situated midway between the pectorals and the 
anal, and have their rays a little coarser, as well as distinctly articulated. The 
dorsal fin, pretty well seen in two specimens, is considerably developed, but 
has not the usual triangular acuminate shape prevalent among Palseoniscidse ; 
on the contrary, its apex is rounded off, and its posterior rays are proportionally 
somewhat longer than is ordinarily the case. Its rays are tolerably coarse for 
the small size of the fish, bifurcating once towards their extremities, and divided 
by somewhat distant transverse articulations. The complete contour of the 
anal fin is not shown in any specimen, but it is seen to commence opposite the 
middle of the dorsal, and to extend close to the lower lobe of the caudal ; its 
rays, so far as they are exhibited, are similar in character to those of the dorsal. 
The caudal, judging from its appearance in three specimens, seems to be not so 
deeply bifurcated as in the more typical representatives of this family, though 
it is very heterocercal and inequilobate ; its rays, similar in general appearance 
to those of the dorsal, are, however, finely and minutely dichotomised towards 
their extremities. I have not, after most careful examination, been able to 
detect any fulcra on the anterior margin of any of the fins. 

Remarks. — Independently of the apparent nakedness of the body, the 
specific novelty of this little fish is fully attested by the peculiar form of 
the dorsal fin, and seemingly also of the caudal, as well as by the absence 
of fin fulcra ; these last considerations being almost of themselves sufficient 
to indicate a new genus. If the body be naked, then not only is a new 
genus requisite, but the occurrence of a Palseoniscoid fish with the 
squamation in a condition almost identical with that of Polyodon, is a most 
interesting fact in connection with certain important structural affinities 



42 RAMSAY H. TRAQU AIR'S 

which, some years ago I pointed out as existing between that living genus 
and the extinct PalaBoniscidae.* 

This is, however, not the first case of the kind which has been recorded, 
for in a paper on the Fauna of the Lower Permian formation of Bohemia, 
by Prof. Anton Fritsch of Prague, I find the following brief notice : — 

" Amblypterus, sp. — Ein kleiner schuppenloser Fisch mit grossen Flossen 
und erhaltenen innern Skeletresten. — Tremosna." t 

No further particulars are here given, but we may look forward with 
pleasure and interest to its full description, as Prof. Fritsch's great work 
on the Amphibia and Fishes of these strata progresses towards completion. 

With regard to the present instance, we may ask, Is it likely that the 
body may have been once clothed all over with scales like those of other 
Paloeoniscidie, but which had been dissolved away or removed by some process 
not at present understood, leaving the delicate bones of the internal skeleton 
uninjured ? As lending some countenance to this view, three specimens of 
other fishes from the present collection may be quoted. 

The first of these is a tail of a very small specimen of Coslacanthus 
Huxleyi, to which I have already alluded (p. 21), and in which no scales are 
visible, though the internal bones are very distinctly preserved. 

The second is the unique specimen, from Coldstream, of the little fish to 
be presently described as Holurus ischypteras, in which the scales are 
not preserved over the whole of the body, though the general form is intact. 

Thirdly, in one of the two specimens of the remarkable form Tarrasius 
problematicus, scales are also invisible on the anterior part of the body. 

With regard to the Coslacanthus, it must be noted that scales are present 
on all the other examples of the species, four in number, though they get 
indistinct towards the tail. The scales of Coslacanthus are always thin and 
delicate, and in shale specimens, at least, they never prevent the more robust 
internal bones being seen ; it is, therefore, perhaps not very wonderful that 
they should have disappeared in the tail of so small a specimen as the one 
referred to. 

Also in the Holurus ischypterus there seems to be evidence of the removal 
of scales by some process of decay, a black film being left on the parts of 
the body where they are absent. But here the process seems also to have 
affected the internal skeleton, which has also almost completely disappeared 
from the parts bare of scales, all that is seen of it being a few oblique 
lines indicating the neural spinous processes, and which are hardly visible 

* " Carboniferous Ganoid Fishes," pt. 1, Paloeontogr. Soc. 1877, pp. 38-40. 

t " Neue Uebersicht der in der Gaskohle und den Kalksteinen der Permformation in Bohrnen 
vorgefundenen Thierreste." — Sitzungsb. der Kbn. bohm., Gesellscb. der Wissenschaften, 21 Marz, 
1879. 



REPORT ON FOSSIL FISHES. 43 

at all, save when the specimen is held in certain lights. The specimen being- 
unique, we had as yet no means of comparing it with others. 

Lastly, as regards the Tarrasius, the specimen in question is also unique, 
in showing the head and anterior part of the body, and further, its obscuration 
by an irreparable film of matrix, renders accurate conclusions really unattain- 
able until better specimens be discovered. (See the descriptions of Holwrus 
ischypterus and Tarrasius problematicus, at pages 66 and 64.) 

But in the case of Phanerosteon, no lateral body scales are seen on any of 
the specimens, four in number,""" in which the body is shown, with the sole 
exception of a few immediately behind the shoulder girdle in one example. 
On the other hand, in all the three which exhibit the tail, the caudal body- 
prolongation is clothed with a set of scales, limited in the very same manner 
in each, and these are as well preserved and as strong as the scales of any 
other Palreoniscid from the same beds. Although the azygous scales in 
front of the dorsal fin are only shown in one example, their being detached 
from such a situation, if not connected with lateral ones, may be readily 
understood. 

If we then consider, finally, that all the other Palasoniscidae which occur in 
the same beds along with the present species have their scales all over the 
body in an excellent state of preservation, the most obvious conclusion seems 
to be that, in this instance, lateral body scales were absent, excepting a few 
immediately behind the shoulder girdle ; the only other alternative being to 
suppose that they were, if present, of an unusually thin and perishable nature. 
In either case the apparent nakedness of the sides of the body, along with the 
other peculiarities noted, seems amply to justify the erection of the genus 
Phanerosteon. 

Position and Locality. — In the Cement-stone group of the Calciferous 
Sandstone series, near Glencartholm, Eskdale. 

Hokums, gen. nov. Traquair. 

Somewhat deeply fusiform ; dorsal fin arising behind the middle of the 
back, not acuminate in front, long based, extending almost to the commence- 
ment of the tail pedicle ; anal fin with a shorter base than the dorsal ; 
caudal strongly heterocercal, but not bilobate, triangular, its rays gradually 
diminishing posteriorly ; pectorals unknown ; ventrals small, abdominal arising 
slightly in front of the dorsal. Scales rhomboidal, sculptured ; a prominent 
row of median scales between the occiput and the commencement of the 
dorsal fin. Teeth small, cylindro-conical. 

The structure of the head is decidedly palseoniscoid, with oblique suspen- 
sorium and wide gape, but none of the specimens afford any evidence of 

* Since the above description was written, a fifth specimen has occurred, with the body as destitute 
of scales as the previous four. 



44 RAMSAY H. TEAQUAIR'S 

the superethmoidal prominence which is so marked a feature in the contour of 
the head in typical PalaBoniscidae. 

Holurus Parki, sp. nov. Traquair. 
PI. III. figs. 9-12. 

Length, 2f inches to apparently over 3 inches ; shape fusiform, and rather 
deep ; greatest depth of body contained about 3^ times, and the length of the 
head a little over 4 times in the total. 

Of the cranial roof bones, the parietals, squamosals, and frontals may be 
readily identified, and these are ornamented externally with sharp and delicate, 
sometimes passing into elongated tubercles. The suspensorium is oblique ; 
the opercular bones seem rather small, and from defective preservation their 
external ornament is not well shown, though in the operculum a few raised 
striae similar to those of the other cranial plates may be observed. The 
maxilla has its upper margin as usual cut away in front for the orbit, but not 
quite so suddenly as in most Palseoniscidse ; its broad post-orbital portion is 
ornamented with delicate ridges running parallel with the superior and posterior 
margins. The mandible is of medium stoutness ; its outer surface shows traces 
of delicate striation. Only very few teeth can, with considerable difficulty, be 
detected ; they are minute and cylindro-conical in shape. 

So far as exhibited, the bones of the shoulder girdle are in every respect 
conformed according to the usual pakeoniscoid type, and are ornamented ex- 
ternally with ridges similar to those of the head bones. 

The body scales are rather small for the size of the fish, rhomboidal, and 
very ornately sculptured with minute and delicate, yet very distinctly marked 
ridges and furrows, whose general pattern on the flank scales (fig. 10) may be 
described as follows : — Below a diagonal running between the anterior, superior, 
and posterior-inferior angles of the scale, their ridges have a nearly horizontal 
direction, parallel with the lower margin, some of the lowest also turning up 
along the anterior margin ; immediately above this diagonal some ridges are 
seen running downwards and backwards parallel with it, while the uppermost 
pass backwards parallel with the upper margin, and then turn down parallel 
with the upper part of the posterior one ; a few denticulations of the posterior 
margin are usually seen about the middle. Further back the denticulations 
disappear, the pattern becomes less marked, the ridges tend to fuse together, 
and the intervening furrows to degenerate into streaks and punctures, till at 
last the minute lozenge-shaped scales on the sides of the powerful caudal body- 
prolongation are nearly smooth. Along the middle line of the back, com- 
mencing near the occiput and extending to the dorsal fin, is a row of large and 
conspicuous median imbricating scales (fig. 12), each emarginate in front, pointed 
behind, and becoming more and more acute as the dorsal fin is approached ; 



REPORT ON FOSSIL FISHES. 45 

externally these scales are sculptured with prominent longitudinal ridges. 
Behind the dorsal fin acutely pointed scales run along the upper margin of the 
caudal body-prolongation in the usual manner. 

No pectoral fin is seen in any of the specimens, and only in one are some 
imperfect remains of a ventral discoverable, this being placed slightly in front 
of the commencement of the dorsal. The dorsal Jin commences behind the 
arch of the back and extends to the commencement of the tail pedicle ; its 
longest rays have only about half the length of the base of the fin ; and as they 
become very gradually elongated in front and remain pretty long behind, a 
peculiarly rounded and proportionally somewhat long-based form of dorsal is 
here produced, which is very different from the high triangular-acuminate 
contour which is prevalent in this family. The anal is somewhat similarly 
shaped, but has a shorter base, for although the termination of its base is 
opposite that of the dorsal, it commences a little further behind. The caudal 
Jin, arising from the lower margin of a powerful body-prolongation, is not 
bifurcated, but assumes a somewhat triangular shape, with the posterior margin 
only gently concave ; its anterior rays being comparatively short, and then 
gradually diminishing posteriorly. The rays of these median fins are nowhere 
seen to dichotomise, but become simply attenuated distally ; they are divided 
by articulations which are distant enough to leave the joints larger then broad ; 
externally they are ganoid, and distinctly striated in the direction of their length. 
No fulcral scales are observable on the anterior margins of any of the fins. 

Remarks. — In its non-bifurcated caudal, and rounded non-acuminated and 
proportionally long-based dorsal fin, this remarkable fish, which I adopt as 
the type of the new genus Holnrus, contradicts the definition of the Palaeonis- 
cidse given by me in the first part of my monograph on Carboniferous Ganoids ; 
and in the want of dichotomisation of the fin-rays, it also differs from all 
hitherto described genera belonging to this family. The apparent want of an 
ethmoidal prominence over the front of the mouth is possibly due to defective 
preservation ; in other respects the structure of the head is so decidedly 
palseoniscoid, that I feel compelled to retain it in this family. After all, the 
differences in the configuration of the fins are of slender importance compared 
with the cranial osteology, and I am inclined to regard it as more convenient, 
for the present, to substitute a more comprehensive definition of the Palaeonisr 
cidse than to institute a new subdivision in these characters alone. Of much 
greater weight are the deviations in the structure of the head, which we shall 
have to consider in connection with the next genus {Canobius.) 

I have pleasure in naming this species after Mr Walter Park, Brooklyn 
Cottage, Langholm, by whose zealous co-operation some of the most interesting 
specimens of the Eskclale fishes were obtained. 

Position and Locality. — Near Glencartholm, Eskclale, in the Cement-stone 
group of the Calciferous Sandstone series. 



4G RAMSAY H. TRAQUAIR'S 

Holurus fulcra tus, sp. nov. Traquair. 
PL III. figs. 13, 14. 

Onty one imperfect specimen of this form has been obtained showing the 
greater part of the body with the dorsal margin and dorsal fin, but deficient in 
the head, the ventral margin, and the fins, except the dorsal. 

Description. — Scales very similar in shape, proportions, and markings to 
those of Holurus Parki, but no denticulations are visible on the posterior 
margins even of the most anteriorly situated flank scales. A row of pointed 
imbricating azygos scales (fig. 14) extends along the middle line of the back, from 
the head to the dorsal fin, and these are much stronger and with fewer and 
coarser ridges than in Holurus Parki. At the commencement of the dorsal 
fin, these median scales pass into a few large and prominent fulcra placed 
along its anterior margin ; the rest of the fin is conformed as in the pre- 
ceding species, extending, as in it, to the commencement of the tail pedicle, but 
the transverse articulations of the rays seem a little more distant. Anteriorly, 
the impressions of a few similar fulcra are seen adpressed to the body, as if 
they had belonged to the pectoral. 

Remarks. — The configuration of the scales and the position and shape of 
the dorsal fin clearly indicate this species as belonging to the same genus as 
Holurus Parki, which it also much resembles in general aspect. Specifically 
it is, however, at once distinguished by the great strength of the median dorsal 
scales, and by the large fulcra in front of the dorsal fin. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 

Canobius, gen. nov. Traquair. 

Body shortly fusiform, rapidly tapering towards the tail ; caudal fin very 
heterocercal, deeply cleft, inequilobate, the upper lobe elongated ; dorsal and 
anal fins short-based, triangular-acuminate, nearly opposite each other, the 
former commencing only very slightly in front of the latter; pectorals and 
ventrals obscure. Suspensorium nearly vertical; snout rounded, slightly pro- 
jecting over the mouth; orbit large, gape small or moderate; dentition un- 
known. Scales rhomboidal, in some cases a row of large imbricating scales 
between the occiput and the origin of the dorsal fin. 

I propose to institute the new genus Canobius for the remarkable little 
fish Canobius Ramsayi, Traq., from the Eskdale beds, a form which to the 
general configuration of a Palaeoniscicl, unites a disposition of the suspensorial 
and opercular apparatus, which is almost identical with the condition of these 
parts in the Platysomid Eurynotus. Here again we have a fish which contra- 
dicts what I once considered an essential character of the Pakeoniscida?. namely, 



REPORT ON FOSSIL FISHES. 47 

the obliquity of the suspensorium ; but "which, according to its other points of 
structure, it would be hard to exclude from that family. I have already ex- 
pressed the opinion (p. 45), that it is meanwhile better to enlarge the definition 
of the group, than to proceed prematurely to break it up into other families. 

It will also be convenient to include under Canobius several other new 
species of Pakeoniscidae, which closely resemble Canobius Ramsayi in external 
form as well as in the direction of the suspensorium, although in certain other 
points of cranial osteology they differ from that species as well as from each 
other. Two of these, namely, Canobius pulchellus and Canobius politus, which 
are rather more typically PaUeoniscid than the others in the configuration of some 
of their head bones, I once thought of forming into another genus ; but, especially 
seeing that the dentition is not yet ascertained in any of these forms, it seems also 
somewhat premature to proceed to the splitting of genera upon these distinctions. 

The generic name is taken from Canobie, the district in which the fossili- 
ferous beds of Glencartholm are situated. 

Canobius Ramsayi, sp. nov. Traquair. 
PL V. figs. 1-4. 

Description. — Length 2^ to 3 inches, shape shortly fusiform, deep in front 
and tapering rapidly towards the tail. The length of the head is contained five 
times, the greatest depth of the body little more than three times in the total. 

The head is short and deep, with a very obtusely rounded snout in front, 
behind which and nearly right over the mouth is a circular orbit of considerable 
size. As far as can be made out, the bones of the cranial roof seem quite 
palaeoniscid in their arrangement, their external surfaces are marked with com- 
paratively coarse flattened corrugations. The suspensorium is nearly vertical, 
being only very slightly inclined backwards ; the posterior margin of the 
opercular flap has a regularly curved semilunar contour. The operculum is 
small, its anterior margin is nearly vertical, but its inferior one so oblique as to 
look as much backwards as downwards, and consequently the posterior 
margin is considerably shorter than the anterior one — the superior being the 
shortest of all. It is succeeded below by an interoperculum of a somewhat 
rhomboidal shape, the acute angles being the posterior-superior and anterior- 
inferior ; its vertical depth is fully as great as that of the operculum, and its 
anterior and posterior margins continue uninteruptedly in the gentle curvature 
of those of that plate. The prseoperculum simulates that of Eurynotus and 
other Platysomidas, being a narrow triangular plate, with acute superior and 
inferior angles, and a very obtuse anterior one : its long posterior margin, 
which fits on to the anterior margins of the operculum and interoperculum 
behind, is gently convex and nearly vertical in position; the other two short 
margins are gently concave, the anterior-superior being the longer, and fitting 

VOL xxx. PART I. II 



48 RAMSAY H. TRAQU AIR'S 

on to the posterior margin of an elongated suborbital, while the shorter 
anterior-inferior one is in contact with the hinder margin of the maxilla. The 
maxilla forms posteriorly a rather broad somewhat rhombic-shaped plate, whose 
anterior angle passes into a narrow process extending on below the orbit. 
The mandible is small, straight, and slender; below it are seen a few branchio- 
stegal rays. Immediately in front of the anterior-superior margin of the prse- 
operculum, and touching the maxilla below, is a narrow, slightly curved sub 
orbital ; and again, in front of this, there is a circlet of narrow ossicles, whose 
number cannot be accurately ascertained, surrounding the entire orbit. The 
orbit is large, and is situated immediately behind the rounded snout, and above 
the anterior part of the maxilla. 

Like the bones of the cranial roof, those of the face are ornamented exter- 
nally with tortuous flattened rugae, except the mandible, which is marked with 
finer and nearly parallel ridges, running from behind forwards, with a slight 
obliquity towards the superior margin. 

No teeth are visible on either jaw. 

The bones of the shoulder girdle are constructed on the usual palaeoniscoid 
type, and ornamented with flattened rugse, like those of the head. 

The scales of the body are arranged as usual in dorso -ventral bands, of 
which 34 may be counted between the shoulder girdle and the commencement 
of the lower lobe of the caudal fin. They are of moderate size, largest on the 
anterior part of the flank, smaller dorsally and posteriorly, and low and narrow 
on the belly. A row of especially large median imbricating scales runs 
along the back from the occiput to the commencement of the dorsal fin. 
These median scales are marked each with a few tolerably well-pronounced 
longitudinal ridges, as are also the imbricating V-scales of the upper caudal 
lobe, but the body scales in general are comparatively smooth, being marked 
only with faint ridges and furrows, proceeding somewhat diagonally from 
before backwards and downwards, which usually stop short before they arrive 
at the posterior margin of the scale ; in many specimens these striae are nearly 
entirely obsolete on the scales below the lateral line. There may also often be 
observed on the flank scales a number of very delicate vertical grooves close to 
and parallel with the anterior margin of the ganoid area. For the most part 
the posterior margins of all the scales are even and entire, denticulations being 
only occasionally and indeed rarely visible. 

The pectoral fin is shown only in one specimen; it is small, and composed of 
numerous delicate rays, which seem to be jointed for a considerable part of 
their length. No ventral is visible in any of the specimens. The dorsal is 
situated far back, so as to be nearly opposite the anal ; both fins are short- 
based, triangular-acuminate in shape, and are composed of delicate, brilliantly 
ganoid and distantly articulated rays. The caudal is very heterocercal, deeply 



REPORT ON FOSSIL FISHES. 49 

cleft and inequilobate, the upper lobe being about twice the length of the lower, 
and nearly equalling one-third of the entire length of the fish ; its delicate rays 
are similar in character to those of the dorsal and anal. Delicate fulcra are 
observable on the anterior margins of all the fins. 

I take the liberty of dedicating this highly interesting species to Professor 
Ramsay, Director of the Geological Survey of Great Britain, to whose kindness, 
and to that of Professor Geikie, I am indebted for the privilege of describing 
this remarkable collection of fossil fishes. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 

Canobius elegantulus, sp. nov. Traquair. 
PI. V. figs. 5-8. 

Description. — Length from 2 to 2f inches ; length of head contained nearly 
five times, greatest depth of the body about 3^ times in the total. Shape 
shortly fusiform, rapidly tapering towards the tail, the upper lobe of which is 
elongated. 

The head is short and deep. 

The cranial roof bones, which are Pakeoniscoid in form and arrangement, are 
marked externally with tolerably sharp, tortuous, and often reticulating ridges. 
The direction of the suspensorium is nearly vertical, the posterior margin of the 
opercular flap evenly rounded. The operculum is a quadrate plate with rounded 
off posterior-superior angle, but its lower margin is not quite so oblique as in the 
last-described species; it is succeeded below by an interoperculum of nearly 
the same size, but having its posterior-inferior angle correspondingly rounded 
off. The praeoperculum is very difficult of detection, but seems to me to be 
represented by a very narrow plate in front of the operculum and interoper- 
culum. In front of this there is, instead of the one long vertical suborbital 
which we saw to exist in Canobius Ramsayi, a chain of three or four short 
ones, in front of which again there is a circle of long, narrow curved ossicles, 
whose number is uncertain, apparently surrounding the entire orbit, which is 
proportionally very large, and seems indeed to occupy almost the entire space 
between the snout and the opercular bones. There is considerable difficulty 
in making out the exact form of the jaw bones. One thing is however certain, 
viz., that the maxilla has not the shape usually found in the Palseoniscidae, 
but has a somewhat triangular form, more resembling that in certain Platy- 
somidse, such as Mesolepis. The gape seems to be small, and the mandible 
delicate ; no teeth can be seen on either jaw. The bones of the face are, like 
those of the cranial roof, sculptured externally with tolerably fine, and occa- 
sionally flattened, tortuous rugae. 



50 RAMSAY H. TRAQUAIR'S 

The bones of the shoulder girdle present nothing calling for special remark ; 
their external surfaces are sculptured in a manner similar to the bones of the head. 

The scales are moderate in size, of the usual rhomboidal shape over the 
body generally, but there is a median row of specially large imbricating ones, of 
a more or less oval shape, extending from the occiput to the origin of the 
dorsal fin, besides the usual V-scales along the upper lobe of the tail. There 
are about thirty oblique dorso- ventral bands of scales from the shoulder girdle 
to the commencement of the lower lobe of the caudal fin. The ganoid area of 
the flank scales shows, in the first place, a few delicate yet sharp vertical 
grooves close to and parallel with the anterior margin, succeeding which, the 
greater part of the exposed surface is sculptured with five or six prominent 
straight ridges running across the scale nearly parallel with the upper and 
lower margins, and ending in sharp jwints on the posterior margin. A very 
similar sculpture pervades the entire squamation, though the corresponding 
ridges on the median scales of the back are somewhat convergent, and the 
minute lozenge-shaped scales of the caudal body-prolongation are nearly smooth. 

I have seen no trace of either pectoral or ventral fins. The dorsal and anal 
fins are nearly opposite each other, the former commencing only a little more 
anteriorly ; both fins are very similar in shape, being short-based and trian- 
gular-acuminate ; each contains about twenty rays, which are delicate, smooth, 
distantly articulated, and dichotomising towards their extremities. The caudal 
is very heterocercal, deeply cleft, and inequilobate, the upper lobe being elon- 
gated; the rays are delicate, smooth, and distantly articulated; the lower lobe 
contains about fourteen rays, but the number of those in the upper one cannot 
be accurately ascertained. 

Remarks. — This very decidedly marked species closely resembles the fore- 
going in size, in the general form of the body and fins, in the shortness of the 
head with its large orbit, and in the direction of the suspensorium, but it may 
at the first glance be distinguished from it by the bold and peculiar sculpture 
of the scales ; the ridges in the head bones are likewise different in character, 
and the dorsal and anal fins seem somewhat more anteriorly placed. In addi- 
tion to these diagnostic characters, an examination of the head reveals certain 
osteological differences, which might easily be considered as indicating a distinc- 
tion of genus. Of these differences the most striking is the form of the maxilla, 
which here assumes a somewhat triangular form, reminding us of that bone in 
Mesolepis, while in Canobius Ramsayi it is not so much modified from the 
ordinary palseoniscid type. Our knowledge of the osteology of the head of 
Canobius elegantulus being, however, still by no means complete, it will, I 
think, be at present more convenient to be satisfied with the more obvious 
resemblances of general configuration, and to leave it provisionally at least in 
the same genus with Canobius Ramsayi. 



REPORT ON FOSSIL FISHES. 51 

Geological Position and Locality. — From the Cement-stone group of the 
Calciferous Sandstone series, near Glencartholm, Eskdale. 

Canobius pulchellus, sp. nov. Traquair. 
PL V. figs. 9-13. 

Description. — The length of one absolutely entire specimen is 2^ inches ; 
that of another, larger, but deficient in the extremity of the upper lobe of the 
caudal fin, is 3^ inches. The length of the latter specimen, when entire, would 
probably amount to \ inch more. 

The length of the head is pretty nearly equal to the greatest depth of the 
body, and is contained slightly more than 4^ times in the total. The shape is 
fusiform, the body being deepest midway between the head and the commence- 
ment of the dorsal fin, and thence tapering rapidly and elegantly towards the 
tail pedicle. 

The cranial roof bones are ornamented with small rounded tubercles, which 
sometimes tend to become elongated, specially on the posterior or parietal 
region. The ethmoidal region forms a rounded projection over the mouth; the 
orbit is large and anteriorly placed. The suspensorium is more oblique than 
in Canobius Ramsayi or Canobius elegantulus, but less so than in typical 
Paloeoniscidse ; the posterior margins of the opercular and interopercular bones 
pass into each other so as to form a continuous gently curved line. The oper- 
culum is a small oblong plate, with rounded inferior margin and posterior- 
inferior angle ; interoperculum, nearly equalling it in size, has its ujDper margin 
correspondingly concave, and its posterior-superior angle slightly produced 
upwards. The prreoperculum cannot be very distinctly made out, but I rather 
suspect that it more resembles that bone in typical Palseoniscidse than in Cano- 
bius Ramsayi. The maxilla is elongated, and its shape is decidedly palreo- 
niscoid, though its broad posterior part is not so suddenly cut away for the 
orbit as in more typical forms ; the mandible is slender and tapering, but 
neither in it nor in the maxilla are any teeth discernible. All the facial bones 
are ornamented with delicate ridges, usually flexuous, though on the mandible 
they are pretty straight, parallel with the inferior margin, and touchiug the 
superior one at acute angles, owing to the tapering shape of the bone. On the 
narrow infra-orbital part of the maxilla the ridges pass into rows of tubercles, 
which pass obliquely downwards and backwards, or, conversely, upwards and 
forwards. The scales are moderate in size ; the median row of scales between 
the head and dorsal fin is rather conspicuous, but not so much so as in 
Canobius Ramsayi and Canobius elegantidus. Taking a flank scale as an 
example, the covered area is narrow, the ganoid one sculptured with closely 
set bold ridges and furrows, forming a pattern which, in its main features, is 
characteristic of a large number of Paheoniscida?. There are first a few 



7v2 RAMSAY H. TRAQU AIR'S 

vertical ridges close to and parallel with the anterior margin, which then turn 
round below and run backwards parallel with the inferior one ; the rest of the 
area is occupied with ridges parallel with the superior and inferior margins, but 
of course directed against the vertical portions of the first-mentioned set. 
Some amount of wavyness is frequently observed in these ridges, and where 
they come to the posterior margin of the scale they end on sharp denticula- 
tions. On other parts of the body, such as the back, belly, and tail, the ridges 
tend to pass into one set which traverse the scales somewhat diagonally from 
before backwards. 

I have not observed either the pectoral or ventral fins. The dorsal and 
anal are nearly opposite each other, the former come only an almost inappreci- 
able distance in advance of the latter; both fins are well developed, short-based, 
triangular-acuminate, composed of tolerably stout smooth rays, which dichoto- 
mise towards their extremities, and are divided by moderately distant articula- 
tions. The caudal is deeply cleft, very heterocercal and equilobate, the upper 
lobe appearing produced ; its rays are similar in character to those of the dorsal 
and anal. Very distinct fulcra are observable along the anterior margins of all 
the fins. 

Several examples have occurred of what seems to me to be only a variety of 
the above described form, the only appreciable difference being in the more 
delicate markings on the scales. 

Remarks. — I know of no previously described fish with which the present 
species can be confounded. In general contour it resembles Canobius 
Ramsayi and Canobius elegantulus, but it may be at once distinguished from 
both by its scale-markings as well as by the more typically palseoniscoid con- 
figuration of its facial bones. In the configuration of the opercular bones and 
the direction of the suspensorium, a condition is presented which is somewhat 
intermediate between that in Canobius Ramsayi and in ordinary Palseonis- 
ciclae, and which, as I have already mentioned, at first nearly induced me to 
institute a separate genus for this and the following species, but considering 
that so much still remains to be learned concerning the more minute characters 
of these small fishes, it is perhaps better to avoid premature multiplication 
of genera by including them provisionally in Canobius, to which they certainly 
bear a greater general resemblance than to any other genus. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 



REPORT ON FOSSIL FISHES. 53 

Canobius politus, sp. nov. Traquair. 
PL V. figs. 14-16. 

Description. — The specimens of this form, none of which have the caudal 
fin perfectly preserved, represent a small fish of from 2^ to 3 inches in length, 
and of somewhat deeply fusiform proportions, the dorsal and ventral margins 
being pretty evenly and elegantly curved. The greatest depth of the body, 
midway between the head and the dorsal fin, is f inch to 1 inch ; the length of 
the head is contained approximately twice in the distance between the tip of 
the snout and the commencement of the dorsal tin, and thrice as far as the 
commencement of the caudal. 

The cranial roof bones are ornamented with comparatively coarse ridges, 
frequently, and in some specimens more than others becoming broken up into 
rounded or elongated tubercles. The snout forms a rounded prominence over 
the mouth, and behind it is placed the orbit, of considerable size. The suspen- 
sorium is only very slightly oblique in its direction ; the operculum and inter- 
operculum are nearly of equal size, and where their external ornament is seen 
it consists of more or less concentric ridges. The shape of the praioperculum 
cannot be made out. The maxilla apparently resembles that of the preceding 
species in form, having a short broad posterior portion, passing into a narrow 
tapering process, which runs forward below the orbit. The mandible is short, 
stout, and straight, and ornamented with longitudinal and oblique ridges, which 
are somewhat finer than those on most of the other bones of the head; on its 
margin several minute sharp conical teeth may be distinguished. 

The scales are of moderate size, largest and least oblique on the front of the 
flank, and diminishing in size posteriorly and towards the dorsal and ventral 
margins. Along the belly, between the pectoral and anal fins, the scales are 
also low and narrow; but those along the middle line of the back are not 
specially large or prominent, excepting a few just in front of the dorsal fin. 
Over nearly the whole of the body the scales are nearly absolutely smooth on 
their exposed surfaces, only on the back, near the middle line, do we observe a 
few grooved striations; and the flank scales of some specimens show some faint 
indications of obsolete ridges, passing with a slight obliquity from before back- 
wards and downwards. The posterior margins of the scales of the side of the 
body are, as far back as the tail pedicle, marked with tolerably well-marked 
denticulations. 

In one specimen a small pectoral fin is visible, but unfortunately its state of 
preservation is not such as to render a minute description warrantable ; 
remains of the ventral are also seen midway between the pectoral and anal. 
The dorsal fin commences only very slightly in front of the anal ; both are of 
the usual acuminate form, with tolerably delicate rays, which are smooth, 



54 RAMSAY H. TRAQUAIR'S 

distantly articulated, and dichotomising towards their terminations ; well- 
developed fulcra are observable along their anterior margins. Only a small 
part of the caudal fin is present in one example, the rays being similar in 
character to those of the dorsal and anal. 

Remarks. — This species is evidently very closely allied to the preceding, 
from which it may, however, be at once distinguished by the smoothness of 
the scales, and by the greater coarseness of the ornament on the cranial roof 
bones, which moreover always partakes more or less of a ridged character ; the 
suspensoriuin seems also slightly more vertical in its direction. Both species 
are referred only provisionally to the genus Canobius. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 

Family Platysomid^e. 

(See Traquair, Trans. Roy. Soc. Edin. vol. xxix. 1880, p. 343.) 

Eurynotus, Agassiz, 1835. 
(Agassiz, Poiss. Foss, vol. ii. pt. 2, p. 153.) 

Eurynotus crenatus, Agassiz. 

A scale indistinguishable from one of the flank scales of Eurynotus crenatus, 
so common a fish in the Calciferous Sandstone series of Edinburghshire and 
Fifeshire, occurs on a small portion of shale from Tweeden Burn, Liddisdale. 

Eurynotus (?) aprion, sp. no v. Traquair. 

PI. V. fig. 20. 

Only a few disjointed scales. One of these, a typical flank scale, measures 
i inch in height by somewhat less in breadth, and closely resembles, in general 
form, a flank scale of Eurynotus crenatus. The well-marked anterior and over- 
lapped area is very distinctly marked off' from the posterior exposed one, which 
is rhomboidal, the acute angles being the posterior- superior and anterior-inferior; 
the ornament consists of tranverse, sometimes oblique, furrows, which are deeply 
marked anteriorly, but fade away towards the middle of the scale, where they 
are replaced by scattered punctures; the posterior margin is quite entire, and 
without any trace of serration or fimbrication. A similar character of ornament 
is displayed by smaller and more regularly rhomboidal scales, which evidently 
belonged to a part of the fish nearer the tail. 



REPORT ON FOSSIL FISHES. 55 

Remarks. — Evidently specifically new, these scales are doubtful as to genus, 
and I only refer them provisionally to Eurynotus on account of their general 
contour and aspect. 

Geological Position and Locality. — Tweeden Burn, Liddesdale, in the Cement- 
stone group of the Calciferous Sandstone series. Identical scales are seen on 
a portion of grey arenaceous shale from the Lower Carboniferous of Moyhee- 
land, Draperstown, Ireland, in the " Griffith Collection " belonging to the 
Science and Art Museum, Dublin, for an opportunity of examining which I am 
indebted to the courtesy of Dr Steele, Director, and Dr Carte, Keeper of the 
Natural History Department of that Institution. 

Eurynotus, sp. indet. 

Among the specimens from Glencartholm is a small Eurynotus, deficient in 
the head and fore part of the body, as well as in the fins, except the dorsal 
and ventral, which are also somewhat imperfectly preserved. The scales are 
striato-punctate, and sharply serrated posteriorly. From the information 
afforded by this specimen, I hesitate either to pronounce it as new, or to 
identify it with any previously described species. 

Wardichthys, Traquair, 1875. 

(Traquair, Ann. and Mag. Nat. Hist. (4), vol. xv. 1875, p. 262; Trans. Roy. Soc. Edin. 

vol. xxix. 1880, p. 361.) 

Wardichthys cyclosoma (?), Traquair. 
PL V. fig. 21. 

A few scales from Tweeden Burn, Liddesdale, and a small fragment of a 
fish from Glencartholm, Eskdale, display characters which I cannot at present 
distinguish from those of Wardichthys cyclosoma, from the Lower Carboniferous 
shales near Newhaven; in consequence, however, of the scantiness of these 
remains, I have appended a query to this determination, so far as the species is 
concerned. The fragment from Eskdale shows merely a small portion of the 
back and shoulder. 

[From Glencartholm there is also a specimen of what is apparently a new 
Platysomid fish, and which may possibly belong to the genus Wardichthys, or 
to some undescribed genus, but its state of preservation is so imperfect that, 
for the present, I abstain from bestowing a name upon it. The specimen 
wants both head and fins, though a portion of the caudal body-prolongation is 
preserved; it measures 3^ inches in length, by If in depth. The shape is 
VOL. xxx. PART I. i 



56 RAMSAY H. TRAQU AIR'S 

more fusiform, less deep and circular than in Wardichthys cyclosoma ; the tail 
pedicle is proportionally strong. The scales of the body, where their surface 
is preserved, are ornamented with coarse, irregular, tuberculo-corrugate 
sculpture ; but on the tail pedicle and caudal body-prolongation their markings 
consist of comparatively delicate, wavy, and more or less diagonal furrows and 
ridges.] 

Cheirodopsis, gen. nov. Traquair. 

Body deep, rounded ; dorsal fin arising behind the arch of the back. 
Scales very narrow. Cranial osteology and dentition as in Cheirodus 
(Amphicentrum, Young). 

The striking difference in the contour of the body, caused by the absence of 
the dorsal, and probably also of the ventral peak, with the greater shortness of 
the dorsal fin, is sufficient to differentiate this genus from Cheirodus, M'Coy, 
to which, in cranial structure, it is most intimately allied. (See the author's 
description of Cheirodus in Trans. Roy. Soc. Edin. vol. xxix. 1880, p. 363). 

Cheirodopsis Geikiei, sp. nov. Traquair. 
PI. V. figs. 17-19. 

Description.— Two specimens of this very interesting form have occurred. 
The first consists of a pretty well preserved head, with the greater part of the 
body and the commencement of the dorsal fin, and when entire probably did not 
exceed 3£ inches in length. The second (fig. 17) is considerably disjointed, but 
represents a somewhat larger fish. So far as it is revealed by the more perfect 
of the two examples, the shape of the fish seems to have been deep and 
rounded, with a very large head compared with the size of the body ; but the 
absence of the posterior part of the specimen renders it impossible to lay 
down any proportional measurements. 

The contour of the head slopes first gently, then, forming an obtuse rounded 
angle above and in front of the orbit, steeply downwards and forwards 
towards the snout ; but the last named part not being preserved, it is impos- 
sible to say whether the prsemaxilla formed the beak-shaped prominence seen 
in Cheirodus. Where the outer surface of any of the cranial roof bones is 
visible, it is seen to be brilliantly ganoid, and ornamented by tolerably coarse, 
tortuous, and reticulating corrugations. 

Judging from the position of the opercular bones, the direction of the 
hyomandibular suspensorium was nearly vertical, or with a slight forward 
inclination. The operculum is not so high as the interoperculum, but both are 
higher than broad ; in form they resemble pretty closely the corresponding 
plates in Cheirodus. Externally they are ornamented with tolerably coarse 
rugae and tubercles ; a diagonal line drawn from the anterior-superior to the 



REPORT ON FOSSIL FISHES. 57 

posterior-inferior angle of the operculum ; and again, turning from the 
posterior-superior to the anterior-inferior angle of the interoperculum, divides 
each plate into two diagonal halves ; behind this line the ornamentation is in 
each case tubercular, while in front it consists of sub-parallel anastomosing 
ridges, running mainly in a horizontal direction. In front of these two plates 
is the interoperculum, shaped as in Cheirodus, but here seen only from its inner 
aspect ; a small narrow additional plate seems to be intercalcated between it 
and the upper part of the anterior margin of the operculum. The orbit is 
placed pretty high up on the head, and right over the middle of the mouth ; a 
narrow sickle-shaped suborbital is seen bounding it behind. 

The mandible is short, deep, and stout, though pointed, beak-like, in front ; 
where the upper margin of its dentary element is seen, it is thin and edentulous. 
The external surface of the mandible is shown in the second and larger speci- 
men, both dentary and angular elements being sculptured with coarse flattened 
rugae, sometimes anastomosing and interrupted, subparallel, and running 
obliquely from above and behind downwards and forwards. The impression of 
the outer aspect of the maxilla is seen in the counterpart of the same sj>ecimen ; 
it is of the triangular shape seen in Eurynotus and Cheirodus, and is ornamented 
by rugae similar in character to those of the mandible, but parallel to the poste- 
rior margin, and nearly vertical to the inferior or oral one. Its free or oral 
margin is sharp and edentulous as in Cheirodus, but probably enough, as in that 
genus, there was a supra-marginal band of tooth tubercles in the inner surface. 

In the smaller specimen an excellent view is afforded of the inner or oral 
aspect of the pterygo-quadrate apparatus, which is conformed exactly as in 
Cheirodus. The pterygoid element, seen also isolated in the larger example, is 
somewhat oval in shape, convex below and internally; its internal surface 
shows a narrow band of small shining tubercles, while the lower margin displays 
two strong ridges, converging behind, and anteriorly carrying each a row of 
tooth tubercles, which are covered with a brilliant coating of enamel, and are 
more closely set than in Cheirodus granulosus. 

Of the bones of the shoulder girdle the only one visible is the clavicle, 
whose form corresponds with that in Cheirodus and other Platysomid fishes. 

The scales (fig. 19) are high and narrow, indeed remarkable for the narrowness 
of the exposed area, which is covered with a peculiar coarse tuberculo-corrugate 
ornamentation, which forms prominent serrations at the hinder margin. The 
articular spine is well marked, as is also the receiving fossette of the internal 
surface, but the vertical keel or so-called " scale rib " is broad, and only very 
slightly prominent. 

The dorsal fin, as seen in the smaller specimen (fig. 17), commences behind 
the summit of the gently rounded line of the back; unfortunately the specimen is 
so cut off that the free margin of the fin is absent, though a considerable 



58 RAMSAY H. TRAQUAIR'S 

portion of the anterior margin, as well as of the base, is visible. The rays, very 
short at first, become rapidly elongated; they are slender and tolerably distantly 
articulated. A small portion of this fin is also seen in the second specimen, 
and here the presence of large and prominent fulcra along the anterior margin 
is very distinctly exhibited. 

The collection contains also a fragment of a tail, which has certainly belonged 
to the counterpart of the last named specimen, the parts between it and the head 
having been lost. This displays part of the upper lobe of the strongly hetero- 
cercal caudal fin, with the posterior fringe-like extremity of the dorsal fin, the 
latter stopping short at the pedicle. The scales of what remains of the tail pedicle, 
and of the caudal body-prolongation, display the nearlyobsolete remains of an 
ornamentation similar in character to that which occurs in those of the body. 

Remarks. — The rounded non-peaked form of the body, with the evidently 
much shorter dorsal fin, seems quite sufficient ground for erecting this remark- 
able new form into a genus distinct from Cheirodus. 

I have much pleasure in naming this species after Professor Geikie, to 
whom I am indebted, on this as on other occasions, for so much kind and 
friendly assistance. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 

Platysomus, Agassiz, 1835. 

(Agassiz, "Poissous Fossiles," vol. ii. pt. 1 ; Young, Quar. Journ. Geol. Soc. 1866; Traquair, 
Trans. Roy. Soc. Edin. xxix. 1880, p. 368.) 

Platysomus superbus, sp. nov. Traquair. 

PI. VI. 
Of several specimens of this splendid fish, one which is remarkably perfect 
may be selected for description. Its principal measurements are as follows : — 

From the tip of the snout to opposite the termination of 

the caudal lobe, ........ 6 inches. 

From the tip of the snout to the bifurcation of the caudal 
fin, .......... 

From the tip of the snout to opposite commencement of 
anal fin, ......... 

From the tip of snout to opposite commencement of ven- 
tral fin, 

Greatest depth of the body from the commencement of the 
dorsal fin, at the highest point of the back, to the 
origin of the ventrals, ...... 



H 


>> 


n 


>> 


2i 


>> 


4f 


» 



REPORT ON FOSSIL FISHES. 59 

The form of the body is therefore very deep ; the back is rounded (though 
with a slight angle at the origin of the dorsal fin), and enormously gibbous ; 
the ventral line is nearly straight as far as the origin of the anal fin, where it is 
obtusely angulated, and slopes upwards to the commencement of the tail 
pedicle. 

So far as the osteology of the head is decipherable, it conforms to the type 
characteristic of this genus ; the cranial bones are ornamented externally with 
delicate, close, wavy subparallel striae, occasionally passing into minute tuber- 
cles. Very distinct imprints of teeth are seen upon the mandible, clearly show- 
ing that these were minute, cylindro-conical, slightly enlarged towards the 
apex, then bluntly pointed. 

The shoulder girdle presents nothing specially worthy of remark. 

The body scales are of moderate size, becoming indeed rather small towards 
the dorsal and ventral margins and the caudal extremity, where, as usual, they 
are also more equilateral. A typical scale from the flank, just behind the head, 
is high and narrow, with well-marked articular spine, and strong internal mar- 
ginal rib or keel. The covered area is narrow ; the exposed one rhombic, with 
very acute anterior-inferior and posterior-superior angles, and is ornamented 
with fine vertical striae, about twelve in the space of £ inch, perfectly parallel, 
and hardly ever bifurcating or intercalated. On the scales further back, and 
towards the margins, especially the ventral one, these striaa often become more 
irregular and wavy, while bifurcation and intercalation very commonly occur. 

By careful working out on the counterpart immediately behind the lower 
part of the clavicle, I succeeded in uncovering a considerable part of the pec- 
toral fin, but not in displaying its perfect contour. Its length is 1 T V inch, being 
greater than the distance between its origin and that of the ventral. 

A little in front of the origin of the anal fin a well-developed ventral is 
exhibited ; it is one inch in length, short-based, and acuminate in form, and is 
composed of numerous rays, which are tolerably closely articulated, and dicho- 
tomise towards their terminations. 

The dorsal fin is remarkable for the large size which it attains both from the 
length of its base and of its rays. It commences at the culminating point of 
the back, slightly in front of the origin of the ventrals, and forms a deep fringe 
extending to the tail pedicle. Its most anterior rays are very short, but they 
rapidly elongate till a length of 1^ inch is attained at the apex, behind which 
the contour of the fin again falls away somewhat, and passes back tolerably 
parallel with the base. The length of the rays in the posterior part is § inch, 
but from the broken up appearance of their extremities, both here and towards 
the apex, it is evident that the full depth of the fin is not exhibited in the 
specimen, a conclusion amply borne out by a fine fragment of a smaller specimen 
to which I shall presently refer. The very numerous dorsal fin rays are ganoid 



60 KAMSAY H. TRAQUAIR'S 

externally, beautifully striated in the direction of their length, with fine straight 
ridges ; anteriorly the transverse articulations are distant, forming joints which 
are considerably longer than broad ; posteriorly they become closer, and the 
joints nearly square, though even here the articulations are more distant 
towards the extremities of the rays. 

The anal fin is in a better state of preservation ; its base, commencing 
f inch behind the origin of the ventral, and extending to the tail pedicle, attains 
only one-half the length of that of the dorsal opposite. Anteriorly it is acu- 
minate, the rays rapidly elongating until a length of 1^ inch is attained at the 
eleventh, whence the contour of the fin again falls away, the posterior part 
being fringe-like, and with rays of about ^ inch in length. The rays are similar 
in character to those of the dorsal, being finely striated longitudinally, and 
having their transverse articulations distant in the anterior and close in the 
posterior rays ; they are also seen to dichotomise towards their extremities, 
while the anterior margin of the fin is set with very distinct fulcra. 

The caudal is of moderate dimensions, heterocercal, and deeply cleft, but in 
this specimen the lower lobe is deficient towards the apex. The rays of the 
lower lobe are pretty stout, divided by closer articulations than those of the 
dorsal, and are rather punctured than striated, although strige appear as we 
pass to the upper division of the fin ; the anterior margin is distinctly fulcrated. 
The rays of the upper lobe are short and delicate, with rather close articulations, 
which, however, still leave the joints rather longer than broad. On the scales 
of this caudal body-prolongation regular striation has disappeared, the orna- 
ment being now reduced to indented furrows and punctures. 

There are two fragmentary specimens in the collection, of which especial 
notice must be taken. One of these represents a portion of the back with the 
upper part of the head of a considerably smaller example than that last de- 
scribed, and showing the anterior part of the dorsal fin in a perfect condition. 
We have here a clear demonstration of the very large development of this fin, 
its longest rays, forming the apex, being larger by one-eighth than the distance 
between the commencement of the fin and the posterior margin of the parietal 
bone. Unfortunately, the hinder part of the fin is not included in the specimen, 
but it is evident that its free margin does not fall away behind the apex, as in 
the case of the anal, and that, consequently, the contour is not so acuminated. 
The anterior margin is distinctly fulcrated. 

In another fragment the entire caudal and anal fins are shown with some 
remains of the ventral. The two latter yield no information beyond what may 
be learned from the first specimen, but the two lobes of the caudal are seen to 
be, as nearly as possible, of equal length. The caudal body-prolongation is, as 
is characteristic of this genus, weak and slender, but its scales are traceable to 
the extremity of the upper lobe. 



REPORT ON FOSSIL FISHES. 61 

Remarks. — This large and beautiful Platysomus, to which I have applied the 
specific name superbus, cannot possibly be confounded with any species pre- 
viously described. Its salient features are — the great gibbosity of the back, 
the great depth of the dorsal fin, whose base is equal to twice the length of 
that of the anal. The scales have a delicate striation, somewhat similar to that 
of the Permian Platysomus striatus, but their exposed arese are much more 
acutely rhomboidal, and the two species are furthermore widely separated by 
the form of the body and of the dorsal and anal fins. 

It is interesting to find in a specimen, which in all respects is an undoubted 
Platysomus, so clear a demonstration of a large and well-developed ventral fin, 
as well as of slender styliform teeth in the jaw. 

Platysomus has not hitherto been found in so low a horizon of the Carboni- 
ferous system. 

Geological Position and Locality. — Near Glencartholm, Eskdale, in the 
Cement-stone group of the Calciferous Sandstone series. 



Of Uncertain Subordinal Position. 

Family Tarrasiid^e. 

Scales rhombic, minute, shagreen-like. Notochord persistent. Neural and 
haemal arches and spines well ossified ; slender interspinous bones penetrate 
between the extremities of the vertebral spines as in teleostean fishes. A long 
dorsal fin composed of closely set jointed rays. 

Tarrasius, gen. nov. Traquair. 

Characters of the Family. — A fragment of a small fish, found by Mr 
Macconochie at Tarras Foot, Eskdale, displays, in spite of its imperfect con- 
dition, characters so startlingly novel, and so completely at variance with 
anything hitherto observed in the domain of palaeozoic ichthyology, that I 
feel compelled to institute for its reception not merely a new genus, but 
likewise a new family. With this I associate a specimen from Glencartholm, 
which displays some of the same characters, and which, so far as evidence goes, 
seems also to belong to the same species. The family and generic names are 
taken from the first locality. 



62 RAMSAY H. TRAQUAIR'S 

Tarrasius problematicus, sp. nov. Traquair. 
PL IV. figs 4-6. 

The specimen from Tarras Foot. — This is a fragment (PI. IV. figs. 4, 5) display- 
ing what is evidently the posterior or caudal part of a small fish, cut off both in 
front and behind, and measuring 1^ inch. The shape of the portion of the body 
here shown is gently tapering, the depth in front being ^ inch, and £ inch 
where it is cut off behind ; the dorsal and ventral margins are nearly straight, 
being only very slightly convex. The whole surface is covered with regularly 
arranged, minute, but proportionally thick, rhombic, and apparently non-over- 
lapping scales, each of which shows on its external brilliant surface a small 
shallow depression or sulcus. At the anterior part of the fragment the 
internal skeleton is clearly displayed by the removal of the scales next the 
eye of the observer. There are no vertebral bodies visible, but four neural (?) 
arches are seen, from which proceed obliquely upwards and backwards as 
many neural (?) spines, in front of which two others are seen, whose support- 
ing arches are not included in the specimen. Above these spines comes a 
series of slender interspinous elements, distally enlarged and laterally flattened, 
while proximally they pass for a little way between the extremities of the 
neural (?) spines, after the manner of modern fishes. Appended to the 
extremities of the last described elements, and set at a slight angle, there 
seems to me to be a second set of interspinous elements, minute, short, and 
somewhat hour-glass-shaped, but owing to the minuteness of the parts it is 
not easy to distinguish them accurately from the proximal extremities of the 
suceeding fin-rays. 

The whole of the dorsal (?) margin exhibited in the specimen is bordered 
by a continuous fin, the depth of which is equal to two thirds of that of the 
part of the body to which it is appended. This fin consists of innumerable 
closely set rays, distinctly articulated, and tapering distally to fine points, 
but so far as can be observed, not dichotomising. As in the continuous clorso- 
caudal fin of Lepidosiren and Ceratodus, their direction becomes posteriorly 
more and more oblique, until at the posterior end of the fragment they are in 
fact nearly horizontal. From this there can be hardly a doubt but that it is 
the tail of the fish with which we have to deal, that the caudal fin was diphy- 
cercal, and continuous with the dorsal and anal. 

On the hsemal(?) aspect of the vertebral axis no arches or spines are 
distinctly exposed, but their presence is betrayed by oblique elevations of the 
scaly surface, exactly symmetrical with the spines of the opposite aspect. 
Along the ventral (?) margin also the impressions of a set of interspinous bones 
are seen, exactly corresponding with those which follow on the neural (?) spines 
opposite, so that although the fin itself is unfortunately lost, we may very 



REPORT ON FOSSIL FISHES. 63 

safely assume the presence of one here also, symmetrical with that on the 
opposite aspect. 

As indicated above, I believe this fragment to be the hinder end of a fish 
with continous diphycercal dorso- and ano-caudal fin ; but as the want of the 
head and abdominal parts render it difficult to distinguish with accuracy the 
symmetrical dorsal and ventral margins, and neural and hcemal aspects, I 
have appended queries to these terms where it has been necessary to use them. 

The specimen from Glencartliohn. — This specimen (PI. IV. fig. 6) is 2f inches in 
length, and presents us in the first place with a head, the length of which is \ inch. 
Unfortunately, very little can be made out concerning the cranial structure. 
On the cranial roof two distinct frontal bones are observable, ganoid externally, 
and faintly sculptured with indented lines and punctures. A distinct opercular 
apparatus is seen consisting of broad plates ; but owing to the crushing they 
have undergone, it is impossible to make out the number or shape of the 
individual elements. The suspensorium is not directed backwards as in typical 
PalseoniscidaB, but seems nearly vertical, if not indeed inclined slightly forwards. 
A considerable portion of a stoutish mandible is seen, marked externally with 
delicate longitudinal ridges, while above it is a portion of a maxilla, but no 
teeth are visible on either jaw. There is also an indication of the position of 
the orbit, right over the middle of the mouth. 

Behind the opercular bones, and somewhat overlapped by them, are some 
traces of a pretty strong clavicle. 

The body is almost completely covered up and obscured by an obstinately 
adherent thin layer of matrix, nevertheless, certain parts are seen, though 
faintly, as if through a veil. The body extends back for 2f inches, or six times 
the length of the head before it is cut off by the edge of the stone ; and as it is 
clear that a pretty considerable portion of the caudal extremity is wanting, the 
fish must have presented a somewhat narrow and elongated contour. The 
points of structure here observable are mainly indications of the internal 
skeleton. For nearly an inch behind the head these are very obscure, consist- 
ing principally of an irregular line, with here and there little bits of bone 
shining through, which are probably portions of neural arches ; behind 
this, however, the line of the vertebral axis is very apparent, although the 
associated structures are very much more clearly seen on the neural than on 
the haemal aspect. As in the specimen from Tarras Foot, there is no evidence 
of vertebral bodies, and the notochord may therefore be presumed to have 
been persistent. On the dorsal aspect of the axial line there is, as in the 
former specimen, a series of slender neural spines, inclined obliquely upwards 
and backwards ; they are pointed distally, but proximally they are enlarged 

VOL. XXX. PART I. K 



64 RAMSAY H. TRAQUAIR'S 

and apparently bifurcated so as to form neural arches. Surmounting these, 
there are also very clear indications of a set of slender interspinous bones, 
whose number is at least double that of the supporting neural spines, and 
whose pointed proximal extremities pass a little way down between the ends 
of the latter ; while again, extending from where the fish is cut off behind for 
fully 1£ inch towards the head, there are evident remains of a long fringe-like 
dorsal fin — in my opinion, a continuous dorso- caudal. Most probably it would 
be found to extend still further forwards could the matrix be removed. 

On the haemal aspect of the axis, clear evidences of haemal arches and 
spines symmetrical with the neural ones above may be seen about two inches 
behind the head, and may be traced for half an inch backwards, beyond which 
the spines become hopelessly obscured, and nothing remains distinguishable 
but the arches from which they spring. 

Finally, in the layer of matrix which obscures the hinder end of the 
specimen, and close to where it is cut off by the edge of the stone, are many 
minute rhombic glittering scales ; at one spot, three of them, apparently portions 
of a dorso-ventral band, are seen in opposition. Each of these little scales has 
a central depression or sulcus, and is, in fact, indistinguishable from those which 
cover the body in the specimen from Tarras Foot. 

Remarks. — The first question which arises concerning the two specimens 
described above, is whether or not they belong to the same species, and here 
difficulties are certainly interposed by the imperfect condition of both. It will 
be observed, however, that there is a very exact correspondence between the 
two as regards the structure of the internal skeleton, so far as this is exhibited, 
and in the long median fin, which extends along a margin which, in the 
Glencartholm specimen, is certainly the dorsal one. The few scales which 
are seen near the caudal extremity of the specimen last referred to, are 
certainly identical in form and appearance with those which thickly cover the 
surface in that from Tarras Foot, and this circumstance, along with the corre- 
spondence of the internal skeleton and median fin, has inclined me to consider 
the two as belonging to the same species. But it must also be observed that, 
whereas the scales in No. 1 cover the entire surface of what remains of the 
body, in No. 2 not a vestige of them is seen till near the posterior extremity. 
If the two specimens really represent the same species, we are reduced to sup- 
posing that in No. 2 the scales have either been loosened by decay and removed 
from the anterior parts (a state of matters which, though not impossible, seems 
hard to reconcile with the fact that the bones of the head and the vertebral 
apophyses are undisturbed so far as the film of matrix allows them to be seen) ; 
or that the obnoxious film of matrix hides them from view ; or lastly, that 
scales were originally present only towards the caudal extremity. It must in 
any case be acknowledged that, until more material turns up, the layer of 



REPORT ON FOSSIL FISHES. 65 

matrix by which the details of No. 2 are obscured, forms an insuperable 
obstacle to a thoroughly satisfactory conclusion on the subject. 

But even if we leave the specimen from Glencartholm altogether out of 
consideration, the fragment from Tarras Foot presents us with peculiarities 
which seem to be quite irreconcilable with the characters of any previously 
defined family. We have scales like those of an Acanthodian, but a position 
in the Acanthodidse is contradicted by the structure of the fin and internal 
skeleton. The general shape and the disposition of the median fin reminds us 
of the hinder part of the interesting Dipnoan (1) fish Co?ichopoma gadiforme, 
Kner, from the Lower Permian of Lebach ; but in that form, as in ordinary 
Dipnoi, the neural and hsemal spines articulate with the interspinous bones, 
end to end, and the squamation is altogether different. It certainly bears no 
perceptible affinity to the Palasoniscidas, nor can I assign to it a place in any 
known family, while, until further material may come to light, even its sub- 
ordinal position is altogether problematical. I therefore propose to institute 
for it the new genus Tarrasius and family Tarrasiidse, both names being taken 
from the locality in which the more characteristic specimen was found. 



66 RAMSAY H. TRAQUAIR'S 



APPENDIX. 



The two following species from Berwickshire were included in this Report 
by an oversight, as the specimens were not derived from the district under 
consideration. In order, however, not to delay their publication, I have 
transferred them from the body of the Report to the end, in the form of an 
"Appendix." 

Holurus ischypterus, sp. nov. Traquair. 

PI. III. figs. 15, 16. 

Description. — Length of the only specimen which has occurred, 2^ inches ; 
but if we allow for some deficiency both at the snout in front, and at the 
termination of the tail behind, the original length would probably be at least 
2^ inches. The other measurements are as follows : — 

Greatest depth of body, . . . . . ^ inch. 

Length from the front to commencement of the dorsal fin, 1^ „ 

Do. do. anal „ 1^ „ 

Do. do. caudal „ If „ 

The shape of the fish is therefore somewhat striking; being elegantly 
fusiform, with a large and apparently non-bifurcated caudal fin attached. 

The head is badly preserved, yet enough is seen to indicate its decidedly 
palteoniscoid structure. The suspensorium is oblique, the gape wide ; the 
operculum is ornamented with a few comparatively coarse oblique ridges ; and 
similar ridges, though somewhat more closely arranged, occur on the post- 
temporal and supraclavicular bones. 

Although the outline of the fish is perfectly shown, only two patches of 
scales are preserved on the side of the body ; one of these is situated on the 
anterior part of the flank immediately behind the head ; the other is towards 
the caudal region, below the dorsal fin, and extends on to the prolongation of 
the body in the upper part of the caudal fin. As shown in these patches, the 
body scales are small, rhomboidal, and marked with two to four very prominent 
ridges, running parallel iviili their upper and lower margins, these ridges 
becoming less marked on the caudal region, where the scales are very minute. 
In front of the dorsal fin are seen some very strong pointed median scales, 
which become indistinct when traced forward ; nevertheless it seems probable 



REPORT ON FOSSIL FISHES. 67 

that they were continued as a distinct row as far as the head. Very strong 
imbricating V-scales are seen along the upper margin of the caudal body- 
prolongation, the sides of which are clothed with minute scales of the usual 
acutely lozenge-shaped contour. 

A thin dark film occupies most of the body space where the scales are not 
preserved, and on this, in the region above the lateral line, are seen, especially 
when the specimen is held in certain directions, certain faint oblique lines 
passing in an upward and backward direction, which seem to indicate the 
vertebral spinous processes. 

The pectoral fin is indicated on the counterpart of the specimen by a 
narrow remnant of its post-axial margin, from which it would seem that it 
nearly equalled the head in length. No traces of the ventrals are discoverable. 
The dorsal commences just behind the highest part of the arch of the back, 
and extends to the tail pedicle ; as H. Parki and H. fulcratus, its anterior 
rays become gradually elongated, and remain long posteriorly, so that the 
contour of the fin rises very gradually in front and finishes off behind in a 
rounded flap-like manner. The anal is like the dorsal in general contour, but 
has a shorter base ; for though the two fins terminate opposite each other, the 
former commences a little behind the latter. The caudal is largely developed, 
but unfortunately its termination is not preserved. Nevertheless, so far as we 
can judge, it seems to have been non-bifurcated, and without any distinct 
differentiation into upper and lower lobes, there being no sudden shortening 
of the rays as they proceed onwards towards the extremity ; the caudal body- 
prolongation is powerfully developed. 

The rays of all these median fins are very delicate, closely set, distantly 
articulated, and ivithout any trace of dichotomisation. Very distinct remains 
of strong and powerful long spicules occur at the bases of the anterior margins 
of the fins, which I interpret as largely developed fulcra, which are even more 
out of proportion with the delicate rays which form the expanse of the fin than 
in the case of Holurus fulcratus. 

Remarks. — This strange little fish cannot possibly be confounded specifically 
with any previously described form ; the only question open to discussion is as 
to the genus in which it ought to be placed. The reasons for referring it to 
Holurus are found in the position and shape of the dorsal fin, the non-bifur- 
cation of the caudal, and the want of dichotomisation in the fin rays. The 
strong fulcra ally it somewhat to Holurus fulcratus, but the ornament of the 
scales is very different from that in the other two species which I include in 
this genus. 

I have already referred to the non-preservation of the scales over a large 
part of the body of this unique specimen, a condition which seems in this case 
at least to have been caused by some process of decay, which has left the fins 



68 RAMSAY H. TRAQUAIR'S 

and head bones, as well as some patches of the scales themselves, intact. There 
is certainly no evidence either of the original absence of scales from the bare 
spaces in question, or of their removal by any mechanical process. 

Position and Locality. — In the Cement-stone group of the Calciferous Sand- 
stone series, left bank of River Tweed, near Coldstream Bridge. 



Canobius obscurus, sp. nov. Traquair. 

Description. — Of this there are only a few fragmentary specimens, which 
indicate a fish of from 2 to 2\ inches in length, and resembling, in shape, the 
other species which I have referred to this genus, being shortly fusiform, 
rather deep in front, tapering rapidly towards the tail, with a short blunt 
head, posteriorly placed dorsal fin, and inequilobate deeply cleft caudal. 

The head has its roof bones covered with fins and tolerably distant ridges, 
frequently interrupted, sometimes branched, and mainly running in a longitu- 
dinal direction, save on the ethmoid, where they are transverse. The rest of 
its osteology is very obscure, but the snout is bluntly rounded, and the suspen- 
sorium seems to be nearly vertical, or at least only slightly oblique ; the bones 
of the face, whose outlines cannot be made out, are apparently ornamented 
with ridges similar in character to those of the cranial roof. 

The scales, proportionally smaller than in the foregoing two species, are 
very regularly rhomboidal in shape, and are marked with from three to Jive 
straight flattened ridges, which pass diagonally from above downwards and 
backwards, and terminate in prominent denticulations of the posterior margin. 
I have observed no row of specially large median scales between the head 
and the dorsal fin, but the dorsal margin is in no instance very well pre- 
served. 

One specimen, more perfect than the others, though the head is wanting, 
shows the dorsal and caudal fins. The former is placed far back, and would be 
nearly opposite the anal, were that fin preserved; it is short-based, triangular- 
acuminate, and composed of very delicate rays. The caudal is deeply cleft, 
and judging from its proportions was doubtless very inequilobate, though the 
extremity of the lower lobe and a considerable part of the upper one are lost. 

Remarks. — The comparatively coarse straight diagonal bars across the 
scales distinguish the species from all the others which I have brought under 
the genus Canobius. In general form it resembles the others, especially 
C. Ramsayi, but unfortunately very little is preserved of the structure of the 
head. 

Geological Position and Locality. — Blackadder Water near Dunse, Berwick- 
shire, in the Cement-stone group of the Calciferous Sandstone series. 



REPORT ON FOSSIL FISHES. 



69 



EXPLANATION OF THE PLATES. 



Throughout these figures the same letters apply to the same hones. 



p. Parietal. 


op. Operculum. 


sq. Squamosal or dermal pterotic. 


i.op. Interoperculum. 


/. Frontal. 


p. op. Prseoperculum. 


p.f. Posterior frontal or dermal sphenotic. 


br. Branchiostegal. 


a./. Anterior frontal or dermal ectoethmoidal. 


s.o. Suborbital. 


e. Median superethmoidal. 


c.o. Circumorbital. 


p.mx. Prsemaxilla. 


s. t. Supra-temporal. 


mx. Maxilla. 


n. Nasal opening. 


pt. Pterygoid. 


or. Orbit. 


mpt. Mesopterygoid. 


p.t. Post-temporal. 


h.m. Hyomandibular. 


s.cl. Supra-clavicular. 


ar. Articular. 


cl. Clavicle. 


ag. Angular. 


p.cl. Post-clavicular. 


d. Dentary. 


i.cl. Infra-clavicular. 


sp. Splenial. 





Fig. 


1.- 


Fig. 


2.- 


Fig. 


3.- 


Fig. 


4.- 


Fig. 


5.- 


Fig. 


6. 


Fig. 


7.- 


Fig. 


8.- 


Fig. 


9.- 


Fig. 


10.- 


Fig. 


11.- 


Fig. 


12.- 


Fig. 


13. 


Fig. 


14. 


Fig. 


15. 


Fig. 


16. 


Fig. 


17. 


Fig. 


18. 



Plate I. 

-Coslacanthus Huxleyi, Traquair ; natural size (p. 20). 

-Another specimen, enlarged by one-half. 

-Angular element of the mandible, magnified three diameters. 

-Scales of Ccelacanthus Huxleyi, magnified six diameters. 

-Elonichtliys serratus, Traquair ; natural size (p. 22). 

-Another specimen ; natural size. 

-Scales from the flank of Elonichtliys serratus, magnified six diameters. 

-Scales from the same species, towards the tail, magnified six diameters. 

-Elonichtliys pulcherrimus, Traquair ; natural size (p. 24). 

-Scales from the flank of Elonichtliys pulcherrimus, magnified four diameters. 

-Scales of the same species, further back, magnified four diameters. 

-Surface of dorsal fin rays, magnified four diameters. 

-Restored outline of Rhadinichthys Geikiei, Traquair (p. 25). 

-Scales from the nape of the neck in Rhadinichthys Geikiei, magnified six diameters. 

-Flank scales, magnified six diameters. 

—The same, a less ornate variety, magnified six diameters. 

-Scales situated further back, towards the tail, magnified six diameters. 

—Narrow abdominal scales, magnified six diameters. 



Plate II. 

Fig. 1. — Rhadinichthys Geikiei, Traquair, var. elegantulus ; natural size (p. 27). 
Fig. 2. — Scales of the same, from the nape of the neck, magnified six diameters. 
Fig. 3. — Flank scales of the same, magnified six diameters. 



Fig. 


4- 


Fig- 


5.- 


Fig. 


6. 


Fig. 


7. 


Fig. 


8, 


Fig. 


9, 


Fig. 


10, 


Fig. 


11. 


Fig. 


12, 


Fig. 


13, 


Fig. 


14, 


Fig. 


15, 


Fig. 


16, 


Fig. 


17. 


Fig. 


18. 


Fig. 


19. 


Fig. 


20, 



70 RAMSAY H. TRAQUAIR'S 

-Scales towards the tails, magnified six diameters. 

-Sketch of the bones of the head, enlarged 2\ times. 

■Rhadinichthys delicatuhis, Traquair; natural size (p. 29). 

-Sketch of the head, enlarged 2J- times. 

-Flank scales of the same, magnified six diameters. 

-Scales further hack, magnified six diameters. 

-Rhadinichthys (f) angustulus, Traquair, magnified two diameters (p. 33). 

-Scales from another example of the same species, magnified six diameters. 

-Rhadinichthys Macconochii, Traquair; natural size (p. 30). 

-Scales from the nape of the neck in the same species, magnified six diameters. 

-Flank scales, magnified six diameters. 

-Scales further back, magnified six diameters. 

-Narrow ventral scales, magnified six diameters. 

-Cycloptychius concentricus, Traquair; natural size (p. 37). 

-Scale from the back above the lateral line, magnified six diameters. 

-Flank scales of the same species, magnified six diameters. 

-Scales towards the caudal extremity, magnified six diameters. 

Plate III. 

Fig. 1. — Rhadinichthys (?) fusiformis, Traquair; natural size (p. 34). 

Fig. 2. — Flank scales, magnified six diameters. 

Fig. 3. — Flank Scales, from another specimen, magnified six diameters. 

Fig. 4. — Narrow ventral scales, magnified six diameters. 

Fig. 5. — Scales towards the tail, magnified six diameters. 

Fig. 6. Phanerosteon mirahile, Traquair ; natural size (p. 39). 

Fig. 7. — Sketch of the head of P. mirahile; enlarged 2| times. 

Fig. 8. — Tail-fin of another specimen, enlarged 2J times. 

Fig. 9. — Holurus Parki, Traquair ; natural size (p. 44). 

Fig. 10. — Flank scales of the same species, magnified six diameters. 

Fig. 11. — Scales further back, magnified six diameters. 

Fig. 12. — Median dorsal scales, magnified four diameters. 

Fig. 13. — Holurus fulcratus, Traquair; natural size (p. 46). 

Fig. 14. — Median dorsal scales, magnified four diameters. 

Fig. 15. — Holurus ischypterus, Traquair; natural size (p. 66). 

Fig. 16. — Flank scales, magnified six diameters. 

Plate IV. 

Fig. 1. — Rhadinichthys tuber xulat us, Traquair; natural size (p. 31). 

Fig. 2. — Flank scales, magnified six diameters. 

Fig. 3. — Scales towards the caudal extremity, magnified six diameters. 

Fig. 4. — Tarrasius problematius, Traquair; Tarras Foot, enlarged two diameters (p. 62). 

Fig. 5. — Scales of the same specimen, magnified eight diameters. 

Fig. 6. — Another specimen, probably referable to the same species, from Glencartholm ; natural 
size (p. 63). 

Plate V. 

Fig. 1. — Canohius Ramsayi, Traquair; natural size (p. 47). 

Fig. 2. — Flank scale, magnified six diameters. 

Fig. 3. — Scale further back, magnified six diameters. 

Fig. 4. — Sketch of head of Canohius Ramsay i; magnified two diameters. 

Fig. 5. — Canohius elegantulus, Traquair; natural size (p. 49). 



REPORT ON FOSSIL FISHES. 71 

Fig. 6. — Flank scale, magnified six diameters. 
Fig. 7. — Scale further back, magnified six diameters. 

Fig. 8. — Sketch of bones of the head in another specimen, magnified two diameters. 
Fig. 9. — Canobius pulehellus, Traquair; natural size (p. 51). 
Fig. 10. — Flank scale, magnified six diameters. 
Fig. 1 1 . — Scale further back, magnified six diameters. 
Fig. 12. — Ganobius pulehellus variety (p. 52). 

Fig. 13. — Flank scale from the same specimen, magnified six diameters. 
Fig. 14. — Canobius politus, Traquair; natural size (p. 53). 
Fig. 15. — Flank scale, magnified six diameters. 
Fig. 16. — Scale further back, magnified six diameters. 
Fig. 17. — Cheirodopsis Geikiei, Traquair; natural size (p. 56.) 

Fig. 18. — Pterygoid bone and outer surface of mandible, from another specimen, natural size. 
Fig. 19. — Flank scales, magnified three diameters. 

Fig. 20. — Scales of Eurynotus (?) aprion, Traquair; magnified four diameters (p. 54). 
Fig. 21. — Scales of Wardichihys cyclosoma, Traquair; magnified four diameters. From Tweedeu Burn, 
Liddesdale. 

Plate VI. 

Fig. 1. — Platysomus swperbus, Traquair ; natural size (p. 58.) 

Fig. 2. — Flank scales, magnified three diameters. 

Fig. 3. — Scales towards the dorsal aspect, magnified three diameters. 

Fig. 4. — Scales of tail pedicle, magnified three diameters. 

Fig. 5. — Sculpture of rays of anterior part of dorsal fin, magnified four diameters. 

Fig. 6. — Dentition of mandible, magnified four diameters. 



VOL. XXX. PART I. 



Trans .Roy. Soc'.Edin r 



Vol. XXX, PL 




9. 





:v^>°v. - A '^& 





v 



I.H.& P A.Traquair del 



F.Hutti,Lith r Zam 1 



Trans. Roy- Soc. Edii 



X 



Vol. XXX, PL II. 




R H.i P.A.Traijuair del 



F HuVh, LifiiT Edm* 



Trans. Roy. Soc. Edm T 



Vol XXX. PI. 111. 




RH.lP.ATraquak lei 



F.Hutti.lilhTEain* 



Trans .Roy. Soc. Edin 1 



Vol XXX. PI. IV. 






R.H.iP.A.Traijuaii ad. 



F.Hufli.LitWEain 1 



Trans. Roy Soc. Edin r 



X 



Vol. XXX, PLY. 




14. 





R H. & PA Traquair del 



Trans. Roj. Soc.Edin T 



Vol.XXX.PLVI. 




PAT 







i J-6.'- 



.H *?A.T«^uair iel 



FHuth.LiUi'Edirf" 



(73) 



IV. — On some new Crustaceans from the Lower Carboniferous Rocks of Eskdale 
and Liddesdale. By B. N. Peach, A.R.S.M., F.G.S., of the Geological 
Survey of Scotland. Communicated by Professor Geikie, F.R.S. (Plates 
VII. to X.) 

(Bead 19th July 1880.) 

By the permission of Professor A. C. Ramsay, LL.D., F.R.S., Director- 
General of the Geological Survey of Great Britain and Ireland, and Professor 
Geikie, LL.D., F.R.S., Director of the Geological Survey of Scotland, I have 
been permitted to describe several new Crustaceans which have come under 
my notice in my capacity of Acting Palaeontologist to the Scotch Survey. 
They are from the -cement-stone group of the Calciferous Sandstone series 
of the Scottish border, and, with a very few exceptions, were got from one 
locality on the river Esk, about four miles south of Langholm, in Dum- 
friesshire, and were almost all collected by A. Macconochie, Fossil Collector to 
the Geological Survey of Scotland. They belong to two orders, viz., Phyllo- 
poda and Decapoda. 

I. Phyllopoda. 

The Phyllopods as yet found in the Calciferous Sandstone series have been 
confined to the genera Dithyrocaris, Leaia, and Estheria. Some large cara- 
paces, attributable to Ceratiocaris, have however been obtained from the 
Carboniferous Limestone, but as far as I have been able to make out, no 
body-segments have been described. In the collection above referred to a 
great many specimens occur, which appear to belong to two different species. 

When compared with those found in the Upper Silurian Rocks they differ 
considerably in having the body relatively much larger than the carapace. The 
abdomen appears to be out of all proportion to the carapace if we take such 
well-known forms as C. papilio (Salter), or C. stygius (Salter), as our ideals. 
The tests are not ornamented by the same fine striations as these latter. 

Ceratiocaris scorpioides, spec, no v. (PI. VII. figs. 1 to 1/). Carapace about 
one-third the length of the body exclusive of telson, subovate in form, and pro- 
duced into a blunt snout anteriorly, and posteriorly into a rounded lobe, which 
extends backwards beyond the median line of the posterior margin. The 
anterior,, ventral, and posterior margins are slightly thickened. The dorsal line 
is almost straight, with a slight droop anteriorly. Jaws two in number, and 
placed within the carapace near its anterior ventral margin. They are hollow 
and denticulated, the toothed portion being much thickened. Body long and 

VOL. XXX. PART I. M 



74 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

cylindrical, and made up of seven segments which extend beyond the carapace, 
and at least five which are covered by it. These latter are mere rings, and 
become shorter and shorter anteriorly. All are articulated and movable. The 
uncovered segments increase in length posteriorly. The third segment from 
the end is much the deepest, whence the body tapers each way. The ventral 
margin of this segment always exhibits a deep notch as if a piece were cut out 
of it. The seventh segment is almost twice as long as any of the others, and 
has articulated with it a strong hastate telson which is about equal in length to 
itself. The telson is broad at the base, tapers rapidly for a short distance, and 
then, more gradually, to a fine point with a slight upward curve. It is orna- 
mented with at least five serrated flanges, a dorsal, a ventral, and two lateral 
ones on each side. The lateral flanges on the one side in part of their course 
are connected with each other at intervals, with buttresses or columns, so that 
a pattern is produced in the lateral groove. The telson is flanked by two 
other shorter conical spines, one on each side, which are also articulated to 
the last body -segment and striated longitudinally. Test, smooth or slightly 
wrinkled. Total length from 1 to 2\ inches. 

Observations. — There is a considerable difference among the several speci- 
mens regarding the length and bluntness of the snout, and from none yet seen 
can the bivalve nature of the carapace be established. 

The species is named from a fancied resemblance of the individuals to scor- 
pions, not from any idea of immediate relationship. 

Ceratiocaris elongatus, spec. nov. (PI. VII. figs. 2 to 2/). Carapace about one- 
fourth of the length of the body without the telson, produced into a long snout 
in front and suddenly deepens where the jaws are seen to show through, 
whence it is produced backwards into a rounded lobe which extends consider- 
ably beyond the medial line of the posterior margin. 

Jaws hollow and denticulated, and placed a little less than half-way from the 
tip of the snout to the posterior margin. They occur nearest the ventral side and 
sometimes project beyond the margin. Body-segments seven, uncovered and 
four or five covered by the carapace, all movable, the whole forming a cylin- 
drical body which swells backwards till the third segment from the tail is 
reached, whence it tapers backwards. This segment too has a similar notch to 
that observed on C. scorpioides, as in that species the segments are each one 
larger than its immediately preceding neighbour. The telson is nearly twice as 
long as the last abdominal segment, and is highly oramented. In addition to 
five plain flanges it has in the grooves on each side of the dorsal flange a row 
of minute bosses, which, when magnified, have the appearance of the pine- 
pattern so common on Indian shawls. Lateral spines not observed. Length, 
including telson, 4^ to 8 inches. 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LTDDESDALE. 75 

Observations — It is only necessary to point out where this species differs 
from C. scorpioides. In the first case its size and proportions are different. 
The snout of the carapace is much longer and narrower. It is in the tail 
spine that the greatest difference is found. Both of the above species differ 
from all others described in the enormously developed abdomen. 

Though the side spines of the C. elongatus were not observed, yet the arti- 
cular surface on the segment to which the telson is attached is much broader 
than there is any apparent need for. It is but natural therefore to infer that 
such spines may have dropped off in the interval between the death and the 
interment of the animals. 

That those of C. scorpioides represent the side spines in the silurian species 
there can be little doubt though they have dwindled down so as to be out of all 
proportion to the central one. We had a shadowing of this, however, in C. 
inequalis, Bar. 

Fig. 2b is interesting, as it shows the course of the intestinal canal, which 
appears to be a straight cylindrical tube opening on the ventral margin of the 
last segment near the insertion of the telson. Fortunately the creature is fos- 
silised with the canal distended with food. 

A noticeable feature in both species is that the pivots on which the abdo- 
minal segments move are placed nearer the ventral margin in the hinder than 
in the anterior segments, thus allowing of most play in the joints of the former 
(fig. la). In both species a row of circular pits is observable on the sides of 
the abdominal segments. These may represent the place of attachment of the 
gill feet with which they were probably provided. 

II. Decapoda macrura. 

Among the specimens are several species which differ in no essential respect 
from the Macrura of the present day. These go to swell the genera Anthra- 
palcemon and Palceocrangon, Salter. It should be borne in mind, however, that 
they are not to be considered as being more nearly allied to the genus Palae- 
mon than to any other of our recent Macrura. As well as these there are 
several specimens of one species differing from the above in having their 
thoracic segments free to move on each other, and not covered by the cara- 
pace, which only extends over the cephalic region. These agree generally with 
the American genus Palceocris* of Meek and Worthen, but the species is 
different from their P. typus.\ 

Genus Anthrapakemon, Salter (1861), Quart. Journ. Geol. Soc. Lond., xvii. 
p. 529. 

* Meek and Worthen, 1865, Proc. Acad. Nat. Sci., Philadelphia, p. 48. 
t Meek and Worthen, Ibid., p. 49. 



76 B. N. PEACH ON SOME NEW CRUSTACEANS EROM THE 

Anthrapahvmon Etlieridgii, spec. nov. (PI. VIII. figs. 3 to 3g; spec. char.). 
Carapace subovate, narrowest in front, and separated into two unequal areas by 
the cervical fold. It is strengthened by a marginal thickening, and produced 
anteriorly into a long serrated rostrum. The posterior angles are rounded, and 
the posterior margin slightly concave. It is further ornamented with five ridges, 
a central one which extends back from the apex of the V-shaped cervical fold 
to the posterior margin. In front of the cervical fold it is continued forward 
into the rostrum. On each side of this there is another ridge passing back to 
the cervical groove but not reaching the posterior margin. On the area in front 
each is continued as two oblong bosses placed upon a low mammiform pro- 
tuberance. These are the supports of spines with which the three already 
mentioned ridges were furnished. The remaining two ridges run almost 
parallel with the lateral margins, but are not found in front of the cervical 
groove. Unlike the rest, these bear no traces of spines nor crenulations. 

The rostrum, which is strong and conical, and about one-third of the length 
of the carapace, is ornamented with a central toothed crest and two lateral ser- 
rated flanges, and ends in a sharp point which is slightly bent upwards. 

The eyes are large, reniform, and pedunculated, and placed at the angles 
made by the rostrum with the carapace which are rounded off into sockets. 

The antennules are two, each made up of a propodite of at least three 
joints, broad, and horizontally flattened at the base and tapering forward. The 
last joint of each supports a pair of short tapering many-jointed setae. The 
antennas consist of a jointed propodite on each side, which supports a broad 
denticulated and corrugated basal scale. It also supports a long many-ringed 
lash, which seems thick at its base compared with the size of the animal. 

The walking limbs, which appear to be five on each side, are stout and 
somewhat flattened laterally and directed forward. From what can be seen, 
they are made up of precisely the same elements as an ordinary monodactylate 
limb of a recent macrurous decapod. There is no evidence of any chelate limb. 

The abdomen consists of six segments irrespective of the telson. The first 
two are narrow and highly facetted, allowing of a great deal of play in the 
joints. Seen from above the unfacetted part is very narrow, but it widens out 
laterally till past the pivots, when it expands into broad and rounded pleurae. 
Those of the second segment overlap those of the adjoining ones both ways. 
Behind this the pleurae are pointed, and overlap those of the succeeding seg- 
ment. The third segment is also highly facetted, but is very much broader 
than any other of the abdominal rings. It is divided down the centre of the 
tergum by a depression in which rises a narrow ridge. The fourth and fifth 
are broad, but not divided medially. The sixth has a central ridge, which is 
continued into the telson. On each side it has a joint articulated with it, each 
of which supports a pair of expanded and rounded fin-like appendages, 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 77 

strengthened by a central or marginal thickening and fimbriated near their 
lower extremities. The telson is broad at its base, and tapers rapidly for 
about two-thirds of its length, where it becomes quite narrow. It then 
expands into a small, oblong, fimbriated flap, with which it terminates. At 
the points where the tapering ceases, a pair or perhaps two pairs of short spine- 
like appendages are articulated with it. Length, f inch to 2 inches. 

Observations. — The above characters have been made out from the study of 
over forty specimens, all of which were collected by A. Macconochie from 
Eskdale. The manner in which the animal is preserved shows that it must have 
been broader than deep, for out of the large number in the Survey collection, 
only one is found on its side. All the rest are preserved back upwards. When 
this is the case the three central ridges on the carapace appear but as if 
slightly crenulatecl, and it was only the specimen preserved on its side which 
showed that these were produced into spines which may be seen to be directed 
forwards (fig. 3). This taken together with the forward direction of the walk- 
ing limbs, the overlapping of the pleurae of the second abdominal segment both 
ways, and the overlapping of the pleurae of those behind it only upon those of 
the next in succession, shows that this animal used the great tail-like apparatus 
made up of the telson and the flattened appendages of the sixth segment for 
swimming backwards, which is the mode of progression in the recent Macrura. 

It is impossible from the manner in which the specimens are crushed to 
make out the relative position of the antennules and antennae, and the 
maxillipedes are never recognisable. The characters on which the classifica- 
tion of the recent Macrura so much depend are therefore not reliable in the 
present case. The general symmetry of the parts best preserved has thus to be 
depended on for that purpose. This applies equally to the other Crustaceans 
here described. The thoracic segments in many cases show through the 
carapace in the manner described by Messrs Meek and Worthen, and R. 
Etheridge, junior, in the species described by them. Fig. 3b exhibits the 
endophragmal system of the thorax very completely. The segments are all 
soldered together except the hindermost. It also shows the gill supports ; five 
or six of these are well seen on one side lying in their proper position. 

The only species already described with which this might be confounded is 
the A. Woodwardi of R. Etheridge, junior,* my friend and predecessor in 
office, after whom I propose to call the present species. It is very much like 
it in general symmetry, though a much larger creature. It differs from it in 
the large size of its rostrum, which as well as the three central ridges of the 
carapace is produced into spines. It is considerably different in the abdominal 
segments, the third being much the largest, while in A. Woodwardi the last is 
the greatest. In A. Etheridgii, the telson broadens out at the tip while the 

* Quart. Journ. Geol. Soc. Lond., 1877, vol. xxxvii. p. 872, t. 27 ; vol. xxxv. p. 468, t. 23. 



78 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

other is pointed. They differ also in the nature of the broadened tail-flaps of 
the sixth segment. 

The ornamentation on the carapace and the broad basal scales of the 
antennae distinguish it from A. Gracilis, Meek and Worthen.* The tail too is 
different, being not quite so complicated as in that species. 

From A. Hilliana, Dawson,t the number of ridges on the carapace show it to 
be distinct. 

Anthrapalwmon Parki, nov. spec. (PI. IX. figs. 4 to 4/). Length from three to 
four inches, and breadth about one-third of the length, which is continued down 
to the sixth abdominal segment, where it expands into a remarkably wide tail. 

Carapace two- fifths of whole length. When expanded it forms an irregular 
quadrilateral figure with rounded off corners, and which is a little narrower in 
front. The anterior margin is concave, and gives off a broad leaf-shaped 
rostrum. The posterior margin is also concave. The margins are strengthened 
by a broad thickening band, broadest at the posterior angles. A deep cervical 
groove, or folding in of the test, proceeds from the anterior angles forming an 
angle of 150° on the median line of the back and divides the carapace into two 
unequally sized areas. The posterior and larger is ornamented by seven ridges 
besides the marginal ones. The middle ridge alone is continued back from the 
cervical groove to the posterior margin, the next two on each side proceed from 
the cervical groove, but are lost before reaching it. The two remaining ridges 
are continued back from the cervical groove, and merge into the lateral 
thickened band at the posterior angles. The area in front is divided into 
several raised portions separated by depressions, but none of the ridges are 
continued on to it. One fold of the test overlaps the base of the rostrum, 
which is leaf-shaped, being narrow where it joins the carapace, expanding 
rapidly and then tapering off to a blunt point. The anterior portion is grooved 
medially and droops downwards. Neither the rostrum nor the ridges on the 
carapace bear any trace of spines or bosses. The test throughout is smooth or 
slightly wrinkled. 

The Cephalic Appendages. — The eyes probably large, and set at the angles 
made by the rostrum with the carapace which are rounded into sockets. The 
antennules consist of two pairs of broadened and jointed peduncles, at least 
three joints are seen which are serrated at their bases, each of which supports 
two jointed setae. The antennae are two broad jointed peduncles which give 
off long jointed setas. No broadened basal scale observed. The rest of the 
cephalic and thoracic appendages not observed. The abdomen consists of six 
segments without the telson. These are short in front, becoming gradually 

* Illinois Geol. Survey Report, 1866, vol. i. p. 407, pi. xxxii. fig. 4, a, b, c. 
t GeoL Mag., 1877, vol. iv., new ser., fig. 1, p. 56. 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 79 

longer, the sixth being the longest. The four posterior segments are orna- 
mented by a broad marginal band anteriorly, which sends back an occasional 
buttress which is soon lost in the test, except in the case of the median one on 
the three last segments, which is continued to their posterior margins. Those 
in a line with the mesial and the two main lateral crests of the carapace are 
larger and more pronounced than the rest, and are continued down into the 
telson. As they approach that organ they become ornamented with occasional 
spines which gradually increase in size backwards. All the abdominal seg- 
ments have large pleurae pointed backwards. The appendages on the first five 
segments have not been observed. Those on the sixth consist of a broad joint 
articulated with it at the posterior angle on each side, each of which supports 
a pair of broad swimming flaps. That corresponding to the exopodite is 
strengthened by a strong, narrow, knife-blade-like rachis on its exterior margin. 
Its inner margin is supported by a conical spine which is directed towards the 
point of the knife-blade portion. The inferior and inner margins are broadened 
out into a flap, which is further strengthened by corrugations of the test. The 
endopodite is composed of a fin-like lobe with a central slender spine-like 
thickening, and is corrugated near the margins. The telson, which is broad at 
the base, tapers rapidly and increasingly for about half its length, whence it is 
continued into a long sharp spine and looks like the section of a boy's peg-top. 
At the angles made by what corresponds to the insertion of the peg, two pairs 
of short conical spines are articulated with it. The convexity of this species is 
slight, as it is invariably fossilised with its back upwards. 

Observations. — Besides the above external characters, the specimens in the 
Survey collection show several points in its anatomy. The eyes as seen, fig. 4«, 
seem to have been large, but they are so much crushed that their original form 
cannot be made out. Nothing can be said about the maxillae or maxillipedes, 
though there is little doubt that the confusion in the carapaces of figs. 4# and 
4& is caused by their being crushed through it. Part of the confusion is, doubt- 
less, owing to the hard parts of the stomach and the endophragmal system. 
In fig. 4 the thoracic segments are seen shining through the carapace, and on 
one side the branchial arches are also distinguishable. In the abdominal seg- 
ments the sternal arches are seen to be pressed up through the tergum, fig. 4. 
Their epimera are almost as broad as the basis of the pleurae, but the sternae 
are very narrow, which shows that the segments to which these belong had a 
great deal of play. The tail is enormous compared with the size of the creature, 
and must have been a most effective organ for swimming backwards. As it is 
now found, in many cases the strengthening ridges only are preserved, as in 
figs. 4 and 4«, and these give to the creature a very formidable aspect. 
Fortunately other specimens, as in fig. 4c, show the true nature of them. 
The sharp knife-blade-like portion of the exopodite is ornamented on its 



80 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

exterior margin by five or six strong conical spines, the broken bases of which 
are seen in figs. 4 and 4c, set in hollows to receive them. Two little reniform 
bosses are seen on each of the abdominal segments as well as the telson. The 
tip of the rostrum has not been observed, as it usually buries itself into lower 
strata than the plane in which the body lies. The test is very thin, and 
probably contained very little calcium carbonate, as it is apt to be filled with 
calculi, such as that now found in the common shrimp. Where the test is thin 
these are mere scales ; but in the spines and thickened portions they are 
semi-globular, the rounded part being mammilated. In every case, however, 
they have a central nucleus from which radiations proceed. The above 
remarks are equally applicable to all the Crustacea described in the present 
paper. Sometimes these calculi are sporadic, at other times they fill the whole 
tests of the creatures, forming an irregular polygonal net- work, which destroys 
the character of the test and gives it a granulated appearance. Even in this 
case the nucleus and radiations are observable. Fig. 4g represents a portion of 
the carapace of fig. 4, magnified about four diameters exhibiting these calculi. 
Fig. 4h is part of the test of the common shrimp Grangon vulgaris, affected in a 
similar manner and magnified about ten diameters. I propose to call this species 
after my friend Walter Park, Esq., Langholm, Dumfriesshire. Though A. 
Macconochie was the first discoverer of the species, Mr Park was the finder of 
the specimen from which fig. 4<# is taken, which he not only handed over to 
the Geological Survey, but generously offered to its collection anything new 
which might turn up to his hammer. Dr Traquair was fortunate enough to 
disinter the magnificent specimen represented by fig. 4. The illustrations are 
all natural size. Fig. 4d is an outline drawing of a portion of the carapace 
and the abdomen to show to what a size this species sometimes attained. A 
portion of the right side of a carapace, shown in fig. 4e, must have belonged to 
a still larger individual. 

Anthrapalcemon Traquairii, nov. spec. (PI. X. figs. 5 to 5/). Carapace short 
compared to length of body, subovate, posterior margin concave, anterior mar- 
gin produced into a rostrum as long as the rest of the carapace. The only 
ornament is a broadened margin slightly crenulated. Cervical fold not dis- 
tinguishable from the other wrinkles of the test. Test smooth and exceedingly 
thin. Rostrum broad at the base and gradually tapering to a point, the upper 
line is slightly hollow and near the point is bent downwards. It is strengthened 
and deepened by a central keel beneath. A transverse section just in front of 
the eye would represent a capital T. Though there are several longitudinal 
grooves near its base, there appear to be no spines. 

Cephalic and Thoracic Appendages. — Eyes stalked and placed on each side 
of rostrum, which is hollowed out to form with the margin of the carapace a sort 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 81 

of socket. Antennules consist of two lengthened, apparently jointed propodites, 
each supporting a pair of filamentous whips, which are cylindrical near their 
base, and then become flattened and broadened somewhat before they taper 
rapidly at the tips. Antennae made up of jointed propodites with prodigiously 
large bidenticulated basal scales. Each propodite also supports a long cylin- 
drical many-ringed lash. The remaining cephalic appendages not distinctly 
observable. 

The walking legs, ten in number, are long and thin in their proportions, and 
somewhat flattened laterally. Their penultimate joints are even more so, and, 
as well, are strengthened by longitudinal flanges. The dactylopodites are sharp 
and spinose, and none of the limbs appear to be chelate. 

Abdomen long, and tapering backwards, composed of six segments, the 
last being the longest and much the narrowest. Their only ornament is a 
slightly thickened band along their anterior margins and a fold of the test along 
the median line of the sixth segment. They all possess triangular, pointed 
pleurae directed backwards. Those of the second segment overlap each way, 
the next three only overlapping those immediately in their rear. No appen- 
dages observable except on the sixth segment, which supports at each posterior 
angle a broad joint, to each of which are articulated a pair of expanded flaps. 
The strengthened part of the exopoclite is much broader than in A. Parki. The 
telson is similar to that of A. Parki, but is narrower and more hastate. It also 
supports a pair of spine-like appendages on each side. 

Observations. — This species somewhat resembles A. Parki, but is easily dis_ 
tmguished from it by the general shape being much more elongated and 
tapering backwards, by the absence of the ridges on the carapace, by the long 
rostrum, by the sudden tapering of the body at the spring of the tail, by the 
telson, and by its possessing large basal scales to its antennae. 

Other points in the anatomy of this species can be made out from some of 
the specimens. The pedipalps of the maxillipedes were in all probability very 
long, and extended in front to near the tip of the basal scales of the antennae. 
The walking limbs seem all to have been simple, as in fig. 5«, ten limbs are 
accounted for, and none seems to be much more enlarged than the rest. They 
agree part for part with the hinder limbs of Nephrops Norvegicus, which they 
greatly resemble. In figs. 5a and 5b the coxopodites of five limbs on one side 
are observable attached to the thorax. Among the confusion produced by the 
crushing together of the gills and endophragmal system, exposed by the cara- 
pace being lifted in fig. 5a, one small fragment of the gill, apparently belonging 
to the fifth thoracic segment, is preserved. When magnified it shows a struc- 
ture like that of the gill of Palcemon, fig. 5e. 

The convexity of this species must have been considerable, for the indivi- 
duals are as often found on their sides as with their backs upwards. When 

VOL. XXX. PART I. . N 



82 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

fossilized sideways the limbs are not crushed up into the body, and are more 
likely to be preserved in a state fit for studying (figs. 5a, b, and c). This also 
applies to all the other species. 

I have named this species after my friend Ramsay H. Traquair, M.D., who 
is describing the fishes got from the same beds with the above, and who has 
handed over to the Survey such Crustaceans as he has been enabled to collect 
during his visits to that locality. 

Anthrapalwmon Macconochii, R. Etheridge, Jun ; Antrapalcemon Macconochii, 

Quart. Journ. Geolog. Soc, 1879, vol. xxxv. p. 471, pi. 23, fig. 10, PI. VIII. 

figs. 6 to Qd. 

The description of the carapace of this species (all of it then known) by 
R. Etheridge, Jun., is so complete that it is unnecessary to add to it. Two 
specimens have recently come to light among those since collected by Mr 
Macconochie which exhibit some of the cephalic appendages as well as the 
body segments and telson in place. The study of these entirely confirms 
Mr Etheridge's opinion that the carapace he described belonged to a 
macrurous decapod. 

Cephalic Appendages. — The eyes are large and stalked, and placed on each 
side of the rostrum. The antennules consist of two pairs of short conical many- 
jointed setae, each pair supported upon a single peduncle, two joints of which 
appear beyond the apex of the rostrum. The antennae are composed of broad 
peduncles, each of which supports a lash which is many-ringed. These are very 
thick at the base when the size of the animal is taken into consideration. No 
basal scale observed. 

The abdomen is short compared with the carapace, and made up of six 
segments, the anterior ones being the narrower. The posterior angles of the cara- 
pace project backwards and inwards like horns, and overlap as far as the ante- 
rior margin of the fifth segment, so that all those so confined are necessarily 
narrow. The sixth expands considerably beyond the tips of the horns, and 
supports a joint on either side, to each of which are articulated a pair of broad, 
rounded, and flattened swimming flaps. 

The telson is broad where articulated with the sixth segment, whence it 
tapers rapidly for a little over half its length, and then expands once more into 
a rounded and fimbriated flap. At the narrowest part it has articulated with 
it on each side a pair of small flaps, so that the tail, made up of all the above 
elements, is a most effective paddle. 

Observations. — The tail greatly resembles that of A. gracilis, Meek and Wor- 
then,* but the carapace is sufficient to distinguish it from that species at a glance. 

Locality. — Tweeden Burn, Newcastleton, Liddesdale. 

Horizon. — Cement-stone group, Calciferous Sandstone series. 

* Proceedings, Acad. Nat. Science, Philad., May 1865, p. 80. 



LOWER CARBONIFEROUS ROCKS OE ESKDALE AND LIDDESDALE. 83 

Anihra'palaemon ornatissimus, nov. spec. (PI. VIII. fig. 7). All that is known of 
this is obtained from the portions of two carapaces on one slab of grey cement- 
stone from Larriston Burn, near head of Liddesdale. 

Carapace about half an inch long, subovate, and produced anteriorly into a 
long denticulated rostrum. The posterior angles and margin not observed. 

It has a deep cervical groove, and is ornamented by a thickened margin, 
which bears two or three rows of small tubercles. The medial line of the back 
bears a slight ridge, which passes back from the apex of the cervical groove, 
is lost in the carapace before reaching half way to the posterior margin. It 
bears no tubercles, and does not occur in front of the cervical fold, where its 
place is represented by a large pyriform elevation. Another line of ridge on 
each side passes back from the cervical fold and runs almost parallel with the 
marginal one. This supports two or three rows of tubercles. At the cervical 
fold it bifurcates, and one branch crosses the fold, and merges into the margin 
at the anterior angle, the other branch coalesces with the margin behind the 
fold. The area in front of the cervical fold is further divided into several raised 
portions by deep sulci. The whole test is studded with minute bosses which 
are much the smallest on the parts of the carapace that are not ridged. It is 
from this character that it derives its name. 

Observations. — This species resembles A. Macconochii, R. Etheridge, Jun., but 
is distinguished from it in having its test covered with minute bosses. The central 
ridge is not continued back to the posterior margin, does not appear on the area 
in front of the cervical fold, and does not bear large tubercles. The other 
ridges have a double or treble row of tubercles, and the rostrum is much larger 
and denticulated. 

Locality. — Larriston Burn, Upper Liddesdale. 

Horizon. — Cement-stone group, Lower Carboniferous. 

Anthrapalcemonformosus, nov. spec. (PI. VIII. fig- 8). Carapace quadrilateral, 
little narrower in front than behind. Anterior margin concave, and produced 
into a long-toothed rostrum. Posterior margin concave, and posterior angles 
rounded. Lateral margin bulging, and strengthened by a thickened band, which 
is broadest in front. There is a deep cervical groove. On the greater area 
behind this, only two ridges occur, one on each side, close to, and parallel with, 
the lateral margin. These pass from the cervical groove to the posterior margin ; 
they end in front in a couple of spines, which overlook the groove. With the 
exception of these, they are quite plain. On a raised mound in front of the 
cervical groove a line of four or five separate spines passes from the interior 
angle of the fold into the median line of the rostrum. Two sulci divide this 
mound from two similar mounds, one on each side, each of which supports two 
separate spines. Along these latter a single spine is set behind the cervical 



84 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

groove. As well as the central ridge of spines, the rostrum bears two lateral 
serrations. 

Observations. — This small species somewhat resembles A Etheridgii in shape 
and in being spinose, but the absence of the three central ridges on the carapace 
possessed by that species distinguishes it. It is not likely to be confounded 
with any other described species. 

Locality. — River Esk, 4 miles south of Langholm, Dumfriesshire. 

Horizon. — Cement-stone group, Lower Carboniferous. 

Genus, Palmocrangon, Salter, 1861 ; Uronectes, Salter, Trans. Royal Soc. 

Edinburgh, vol. xxii. p. 394 ; Palmocrangon, Salter, 1861, Quart. Journ. 

Geol. Soc. vol. xvii. p. 533. 

Palmocrangon Eskdalensis, nov. spec. (PI. VIII. figs. 9 to 9z). A shrimp-like 
creature, about 1^ to 2 inches in length. Carapace one-third of the length of 
the body without the appendages. Seen sideways, it is subquadrate, narrow, 
and blunt in front, and produced into rounded lobes posteriorly, which extend 
beyond the posterior margin on the middle line of the back, and overlap the 
pleurae of the first abdominal segment. The only ornament is a slight marginal 
ridge, and the anterior margin is slightly serrated or denticulated. A fold of 
the test which runs almost parallel with and very near to the anterior margin, 
and which supports two or three denticles, may represent the cervical groove. 
The rostrum is not much larger than one of the denticles on the anterior 
margin. Eyes elongated and pedunculated. The antennules are made up of 
long three or four jointed propodites, upon each of which are placed two equally 
long jointed seta?. The antennae are made up of long many-jointed flagellar 
and broad basal scales, the tips of the latter not extending beyond the propo- 
dites of the antennules. Of the other appendages belonging to the cephalo- 
thorax only a doubtful appearance of a maxillipede has been observed, and 
represented in fig. 9c, but its component parts are not traceable. 

The abdomen is cylindrical, and tapers each way from the third segment. 
It is made up of six segments, each slightly modified to suit the region where 
it is placed. The first is a mere narrow ring articulated with the last thoracic 
segment, and has small pleurae. The pleurae of the second expand into 
rounded saddle-flap like lobes, which overlap those of the segments both in 
front of it as well as behind. At the pivot where this one moves upon the first 
segment, the anterior margins of the pleurae are bent forwards at almost a 
right angle till they touch and overlap the posterior lobe of the carapace, so 
that the pleurae of the first segment are entirely hidden. The third, fourth, 
and fifth segments have pointed pleurae directed backwards and overlapping 
their succeeding neighbours. The sixth segment is much longer than any of 
the rest, and forms a conspicuous feature. 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 85 

Abdominal Appendages — The appendages on the first abdominal segments 
seem to be modified for sexual purposes. Figs. 9b and 9c show articulated 
with this a stout longish limb-like appendage, made up of five joints, which 
reminds one somewhat of the corresponding modified limb of the male cray fish, 
Astacus fluvialitis, only it is much larger in proportion. The best preserved case 
of this is shown magnified in fig. 9/ Analogy would incline one to look at 
those other individuals exhibiting the above character as males. The ap- 
pendages on the next four segments appear as short and flattened, the seg- 
mentation not being traceable in the present state of preservation. The 
appendages of the sixth segment are made up of the broad propodites articulated 
with it at its posterior angles, each of which gives rise to a pair of swimming 
flaps with strengthened ridges ; those of the exopodite are blade-shaped and 
slightly curved outwards. The telson is long and hastate, and supports a couple 
of spines on each side. Test smooth, or only slightly pitted in appearance very 
like that of a shrimp. 

Observations. — It will be at once observed, from the enlarged lobes of the 
second abdominal segment and their overlapping each way, that it was here the 
animal doubled itself while swimming, just as in our recent shrimps. Like P. 
socialis, Salter, it often occurs in great numbers. It differs from that species 
in the shape of the carapace and in the pleurae of the first and second 
abdominal segments as well as in the more elongated telson. 

Locality. — River Esk, 4 miles south of Langholm, Dumfriesshire. 

Horizon. — Cement-stone group, Lower Carboniferous. 

Note. — Since the above was written, the Rev. Thomas Brown, F.R.S.E., in 
whose possession the specimens of P. socialis are from which Mr Salter made 
the genus, has kindly allowed me to see them. There is no doubt that our P. 
EsMalensis belongs to the same genus, but there is sufficient difference to 
warrant its being ranked as a separate species. 

Genus Palmocaris, Meek and Worthen, 1865, Proc. Acad. Nat. Sci., 
Philadelphia, 1865, p. 48. 

Palceocaris Scoticus, nov. spec. (PI. X. figs. 10 to 10/*). Body long and 
narrow, and tapering backwards. Length from \ of an inch to 1 inch. The cara- 
pace only extends over the cephalic segments, and measures about two-sevenths 
of the body irrespective of telson and appendages. It is rounded in front, 
and its posterior margin somewhat concave. It is divided into several areas by 
deep sulci. One of these of a V-shape extends slightly back from the anterior 
margin, and has its apex directed backwards. Other two proceed from the 
anterior portion of the lateral margins, and reach the posterior margin after 
having performed crescent-like curves, the convex sides of which are directed 
inwards ; from these latter short grooves inwards proceed, but do not meet on 
the median line. 



86 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

Cephalic Appendages. — No eyes have with certainty been observed. 
Antennules two, each made up of a long propodite of three or four cylindrical 
joints supporting two many-jointed whips, the tips of which have not been 
preserved, so that nothing can be said as to whether the outer or inner be the 
longer. The antennae consist of a shorter and thicker propodite, which 
supports a rounded basal scale and a many-jointed flagellum, which is much 
stouter than those of the antennules, and probably extended much beyond 
them, though only the bases are now preserved. The other cephalic ap- 
pendages have not been observed. 

Thorax and abdomen are composed of similar segments, which are twelve 
in number ; those of the thorax not having the terga and pleura soldered to- 
gether nor confluent with the carapace. All are free to move on each other, 
and are composed exactly like a typical abdominal segment of a decapod. 
Each ring exhibits a tergum flanked by pleurae, which are marked off from it 
by a folding in of the test. These grooves are continued backwards from those 
already mentioned as occurring on the carapace to the anterior lateral margins 
of the telson. The pleurae are pointed, and directed backwards. The body 
tapers gradually from the second thoracic segment. The three last abdominal 
segments are ornamented on the terga with two ridges, situated just within the 
grooves, and these are continued on the telson. No appendages have been with 
any certainty seen on any of the above segments, except on the last abdominal 
one, which supports a joint at each posterior angle. To each of these are 
attached a pair of broadened swimming flaps, the strengthening portions of 
which are usually alone preserved as flattened spines ; those of the exopoclite 
being the longer and more pronounced. The telson is shield-shaped, broad at 
the base and tapering almost to a point, and then suddenly broadening out 
spoon-fashion at the tip. At least one pair of small spines are attached to it 
immediately above its narrowest part. The tail is in character quite that of a 
macrurous decapod. The test is ornamented all over with slight corrugations, 
which, along with a metallic lustre the fossil usually presents, makes the 
slightest fragment recognisable. 

Observations. — This species was found by A. Macconochie at the locality on 
the Esk where the other fossils were procured, and from which about thirty speci- 
mens have been obtained. From the nature of the matrix, which is a sandy and 
calcareous shale, and the fact that the test of even these tiny creatures are infested 
with innumerable calculi of the character already described, they cannot be 
studied as narrowly as might be wished. For instance, although in most of the 
specimens there occur small oblong bosses just in the place where their eyes 
should be, were they decapods, figs. 10-lOcZ, yet the facets of the cornea have been 
looked for in vain. This is unfortunate, as it prevents one from saying with 
certainty that these are the eyes, though there is a strong presumption in favour 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 87 

of their being so. No sessile eyes have been observed on the carapace, neither 
has a trace of anything been observed that could be construed into such. The 
nature of the other cephalic appendages preserved is quite that of the decapods, 
and the same is the case with the tail (see figs. 10-1 Oe). The general appearance 
of the thorax and abdomen much resembles the Isopods, but this is only seem- 
ing. Nothing can be said of the walking feet or those of the first five abdominal 
appendages, as in no specimen yet obtained have these been preserved in a 
state to be studied, though in some cases there is a faint appearance as if of 
something of the kind. 

In the specimen represented by fig. 10b, which is among the few that has 
been fossilised on its side, the appearance of several segments shows through the 
carapace. The lobe of the carapace is wrinkled with lenticular sigmoid mark- 
ings which may possibly represent the gill arches showing through. It also 
overlaps the pleura of the first thoracic segment so as to entirely hide it, which 
lends additional strength to this supposition, as it indicates how free it is. As 
far as the present specimens go to show the affinities of these small crustaceans, 
I am inclined to the belief that they are lowly decapods, somewhat like the 
opossum shrimp (Schizopods), though not necessarily identical with them. For 
comparison, I have reproduced the figures which Messrs Meek and Worthen 
give of their P. typus in the Memoirs of the Geological Survey of Illinois, vol. ii. 
pi. xxxii. Fig. 10 is fossilised back up, as most of ours are. In it thirteen seg- 
ments are shown in the combined thorax and abdomen, but the depressions which 
appear to form the first do not meet in the middle, I am inclined to look upon 
it as belonging to the head, and the sulci as analogous to those seen in the 
head of our specimens. The same may be said of fig. lOg. If that be the case, the 
number of segments in each is the same. The American specimens seem to be 
preserved more in the round than the present ones, hence their generally 
narrower appearance and their not exhibiting the lobes of the carapace. Fig. 
10k is the restored tail of P. typus, alongside which, fig. 10<?, a restoration of 
that of P. Scoticus, from the study of more than twenty specimens, is put for 
comparison. 

As far as I am aware, this is the first species referable to the genus Palceo- 
caris that has been procured from British strata. 

Note 1. — The study of the tails of the macrurous decapods, described in the 
present paper, tends to confirm the opinion of those who hold that the telson 
in the Macrura is only a modified body segment, for all the telsons that are suf- 
ficiently well preserved to retain them have two small swimming flaps or spines 
articulated with them on each side. These appear to be modified appendages 
and probably represent the exopodites and enclopodites of those of the 
other segments. Even the lowly Palceocaris exhibits these articulations. 



88 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

In the interval between the reading the present paper and its printing, I 
have been enabled to study several more Carboniferous Macrura. Some new- 
species of Pala?ocragon, each showing the swimming flaps flanking the telson and 
articulated with it ; also a large suite of fossils of macrurous decapods inter- 
mediate between Palceocaris and Anthrapalwmon, in having only two or three 
of the thoracic segments covered by the carapace, and the sterna of these not 
soldered together, and they all exhibit the spines or flaps on the telson. The 
small spines articulated at the posterior angles of the telsons of our recent 
shrimps and prawns are evidently a survival of what has once been a useful 
character of their more ancient progenitors. It seems a pity to disturb the 
seemingly satisfactory and complete number twenty for the segments of the 
typical crustacean by adding another to make it twenty- one, but the present 
evidence favours the latter number. 

Note 2. — Since the above was written, a paper read by M. P. Brocchi before 
the Geological Society of France, in Nov. 1879, and published in the Bulletin 
of that Society in March 1880, on a " Fossil Crustacean, from the Schistes 
d'Autun" (Upper Carboniferous or Lower Permian), has come before my notice. 
After a careful perusal of his paper, I have not been able to satisfy myself with 
the conclusions M. Brocchi arrives at regarding the systematic position he assigns 
to the genus Palceocaris, Meek and Worthen, which he includes in his new 
division of the Amphipoda, the Nectotelsonides. His reasoning, based upon the 
first thoracic limbs not being modified into prehensile organs, disappears in the 
light of such forms as Anthrapalcemon Traquairii and A. Etheridgii described in 
the above paper, or the recent Palinurus vulgaris, whose limbs are not so modified, 
but in which the decapod characters are undoubted. Such evidence as the 
specimens of P. Scoticus afford are in favour of decapod affinities, and I there- 
fore retain the classification latterly adopted by Messrs Meek and Worthen, 
which M. Brocchi has apparently overlooked.* 

* Memoirs of the Geological Survey of Illinois, vol. iii. p. 552, 1868. 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 89 



EXPLANATION OF PLATES. 



PLATE VII. 



Fig. 1. — Ceratiocaris scorpioides, showing the general form. The carapace is slightly displaced, and 
exposes some of the segments which during life were covered by it. Jaws seen near anterior 
margin of carapace. Nat. size. Locality — River Esk, 4 miles south of Langholm, Dum- 
friesshire. 
Fig. la. — Ceratiocaris scorpioides, showing carapace slightly displaced, jaws, &c. Nat. size. Same 

locality. 
Fig. lb. — Cast of jaws of fig. 1, magnified. 
Fig. \c. — Crushed jaw of fig. la, magnified. 
Fig. Id. — Tail spines of fig. la, magnified. 

Fig le. — Magnified tail spines of another specimen of Ceratiocaris scorpioides, from the same locality. 
Fig. If. — Magnified tail spines of another specimen of Ceratiocaris scorpioides, from the same locality. 
Fig. 2. — Ceratiocaris elongatus, showing all but last joint and tail spines. Carapace displaced so as to 
show some of the segments which it should cover. Jaws seen in situ. Nat. size. Same 
locality. 
Fig. 2a. — Anterior portion of Ceratiocaris elongatus, showing lobe of carapace wrinkled, and notch in 
ventral surface of third segment frontal. Nat. size. Same locality. 
-Ceratiocaris elongatus, showing intestinal canal, Nat. size. Same locality. 
-Showing natural size of last body segment and tail spines of Ceratiocaris elongatus. 
-Portion of tail spine of fig. 2c, magnified, to show ornamentation. 

-Tail of new species of Ceratiocaris from the Upper Silurian (Wenlock) Rocks of Roxburgh- 
shire, for comparison of ornament with fig. 2d. 
-Ornament of 2e magnified. 



PLATE VIII. 

Fig. 3. — Anthrapalamion Etheridgii, magnified 2 diameters. Locality — River Esk, 4 miles south of 

Langholm, Dumfriesshire. 
Fig. 3a. — Anthrapalamion Etheridgii. Nat. size. Same locality. 
Fig. 3b. — Anthrapalaimon Etheridgii, to show endophragmal system and gill supports. Magnified 

2 diameters. Same locality. 

Fig. 3c. — Portion of carapace of Anthrapalcemon Etheridgii, to show the number and nature of its- 

ridges. Nat. size. Same locality. 
Fig. 3o\ — Anthrapalamion Etheridgii, anterior portion of carapace, to show the antennae and their 

basal scales. Magnified about 3 diameters. Same locality. 
Fig. 3e. — Magnified portion of specimen of Anthrapalosmon Etheridgii, to show eyes. Magnified about 

3 diameters. Same locality. 

VOL. XXX. PART I. 



Fig. 


2b. 


Fig. 


2c, 


Fig, 


2d. 


Fig. 


2e. 


Fig. 


2/ 



90 B. N. PEACH ON SOME NEW CRUSTACEANS FROM THE 

Fig. 3/ —Magnified portion of another specimen of Anthrapalcemon Etheridgii, showing antennae, with 

their hasal scales. Magnified 4 diameters. Same locality. 
Fig. 3g. — Anthrapalcemon Etheridgii. Magnified dorsal portion of second abdominal segment, to show 

nature of ornament of test. Same locality. 
Fig. 6. — Anthrapalcemon Macconochii, R. Etheridge, Jun., showing body segments and cephalic 

appendages. Nat. size. Locality — Tweeden Burn, Newcastleton, Liddesdale. 
Fig. 6a. — Anthrapalcemon Macconochii. Nat. size. Same locality as preceding. 
Fig. 65. — Carapace of Anthrapalcemon Macconochii, to show nature of ornament. Nat. size. Same 

locality. 
Fig. 6c. — Anthrapalcemon Macconochii. Magnified anterior portion of fig. 6, showing eyes, antennules, 

and antennae. 
Fig. Qd. — Tail of Anthrapalcemon Macconochii, magnified from fig. 6. 

Fig. 7. — Anthrapalcemon ornatissimus. Nat. size. Locality — Larriston Burn, Upper Liddesdale. 
Fig. 8. — Anthrapalcemon formosus. Nat. size. Locality — River Esk, 4 miles south of Langholm, 

Dumfriesshire. 
Fig. 9. — Polceocrangon Eskdalensis. Nat. size. Same locality. 

Fig. 9a. — Polceocrangon Eslcdalensis, showing eye, antennules and antennae. Nat. size. Same locality. 
Fig. 9b. — Polceocrangon Eslcdalensis, showing antennules and basal scale of antenna, appendages, &c 

Nat. size. Same locality. 
Fig. 9c. — Polceocrangon Eskdalensis, exhibiting same as above. Nat. size. Same locality. 
Fig. 9d. — Polceocrangon Eskdalensis. Nat. size. Same locality. 

Fig. 9e. — Carapace and basal scales of antennae of Polceocrangon Eskdalensis. Nat. size. Same locality. 
Fig. 9/. — Modified appendage of first abdominal segment of Polceocrangon Eskdalensis. Magnified from 

Fig. 9b. 
Fig. 9g. — Last abdominal segment and tail of Polceocrangon Eskdalensis. Nat. size. Same locality. 
Fig. 9A. — Polceocrangon Eskdalensis. Last abdominal segment and tail. Nat. size. Same locality. 
Fig. 9i. — Eye of fig. 9a, to show the crushed facets of cornea. Highly magnified. 
Fig. 9k. — Restoration of tail of Polceocrangon Eskdalensis, from the study of a great number of 

specimens from the above locality. 

PLATE IX. 

Fig 4. — Anthrapalcemon Parki, showing antennules, antennae, rostrum, abdomen, and tail. Nat. size. 

Locality — River Esk, 4 miles south of Langholm, Dumfriesshire. 
Fig. 4a. — Anthrapalcemon Parki, showing carapace, rostrum, eyes, antennae, three abdominal segments, 

and tail. Nat. size. Same locality. 
Fig. 45. — Anthrapalcemon Parki, to show nature of ornament on carapace. Nat. size. Same locality. 
Fig. 4c. — Last abdominal segment and tail of Anthrapalcemon Parki, showing that the spines in the 

other specimens are portions of fin-like lobes, shows spinelets on the outer margin of the 

exopodite and small spines articulated with telson. Nat. size. Same locality. 
Fig. 4d. — Outline of specimen of Anthrapalcemon Parki. Natural size, to show to what a size they 

sometimes attained. Same locality. 
Fig. 4e. — Outline of portion of right margin of carapace of still larger individual. Nat. size. Same 

locality. 
Fig. 4/ — Restored tail of Anthrapalcemon Parki, made out from the study of a large suite of specimens. 
Fig. 4/j. — Portion of test of common shrimp (Crangon vulgaris), highly magnified as a transparent 

object, to show nature of calculi which infest it. 
Fig. 4#. — Portion of carapace of fig. 4, Anthrapalcemon Parki, showing similar calculi in these ancient 

crustaceans, magnified 3 diameters, as an opaque object. 



LOWER CARBONIFEROUS ROCKS OF ESKDALE AND LIDDESDALE. 91 



PLATE X. 

Fig. 5. — Anthrapalcemon Traquairii, showing general proportions. Nat. size. Locality — River Esk, 

4 miles south of Langholm, Dumfriesshire. 
Fig. 5a. — Anthrapalcemon Traquairii, fossilised sideways, shows carapace, rostrum, eyes, antennules, 

antenna?, with basal scales, coxae of walking limbs, walking limbs, gills, and portions of 

six abdominal segments. Nat. size. Same locality. 
Fig. 5b. — Anthrapalcemon Traquairii, exhibits portions of ten walking limbs. Nat. size. Same 

locality. 
Fig. 5c. — Anthrapalozmon Traquairii, showing walking limbs. Nat. size. Same locality. 
Fig. 5d. — Basal scales of antennae of Anthrapalozmon Traquairii. Nat. size. Same locality. 
Fig. 5e. — Highly magnified portion of gill of fig. 5a. 
Fig. 5f. — Restored tail of Anthrapalasmon Traquairii, made out from the study of several specimens 

which are not figured for comparison with that of A. Parki. 
Fig. 10. — Palwocaris Scoticus, magnified 2 \ diameters. Same locality. 
Fig. 10«. — Palceocaris Scoticus, showing antennules, antennae, with basal scale. Magnified 3 diameters. 

Same locality. 
Fig. 106. — Palwocaris Scoticus, fossilised sideways, shows rostrum, eye (?), segments appearing through 

carapace, and overlapping of lobe of carapace over the pleura of first thoracic segment. 

Magnified about 2 diameters. Same locality. 
Fig. 10c. — Head and appendages of Paloeocaris Scoticus, magnified 5 diameters to show rostrum, eyes (?), 

antennules, antennae with basal scale, and folds of carapace. Same locality. 
Fig. lOd. — Another specimen, magnified 5 diameters, showing the same cephalic appendages as fig. 10c. 

Same locality. 
Fig. lOe. — Restoration of tail of Palwocaris Scoticus, from the study of over 20 specimens, to compare 

with fig. 107j. 
Fig. 10/. — Reproduction of figure of Palceocaris typus, of Messrs Meek andWorthen, from the Memoirs of 

the Geological Survey of Illinois, U.S.A., vol. ii. plate 32, for comparison with P. Scoticus. 
Fig. \0g. — Paloeocaris typus. Reproduction of Messrs Meek and Worthen's figure, ibid. 
Fig. 10/i. — Reproduction of Messrs Meek and Worthen's restoration of their Palceocaris typus, ibid., to 

compare with fig. 10c. 



Trans. Royal SocEdm 1 



VolXXX, Plate VII. 




B.'K.Peach.ael. 



FHutti, lift* Elm* 



Trans. Royal Soc.Edirf 



Vol.XXX, Plate VIII. 




EH PeacTi.ael. 



F Huft.Iitii 1 Edm x 



Trans. "Royal Soc.Edin 1 



Vol XXX, Plate IX. 






4 K: 




-/- f 





p 



&y 






v 



4<£. 



4 



# 




4 #A 




4 




^ 



B N.Peach, del. 



FHuth.I.ift* Efo 



Trans. Royal Soc. Ediif 



Vol XXX, Plate X. 




F Huth, liOf ESm' 



(93) 



97 



V. — Gaseous Spectra in Vacuum Tubes, under small Dispersion and at loiv 
Electric Temperature; including an Appendix III., by Prof. Alexander 
S. Herschel, M.A., Newcastle-on-Tyne. By Piazzi Smyth, F.R.S.E., 
and Astronomer Royal for Scotland. (Read July 19, 1880). 

CONTENTS. 

PAGE 

General Introduction, ....... 93 

Practical Commencement described, . . . 96 

On the tables of Twenty Gas- Vacuum Tubes, .... 97 and Appendix I. 

Examination of the observed Quantities, and Elimination of " Im- 
purity" Effects, ....... 

Search for New Lines and their Gaseous Identifications, . . 99 and Appendix II. 

Standard Tables of the Principal Gaseous Lines and Bands, . . 99 

Of Changes with Time and Use, . . . . 100 

On recent observations in Belgium, ..... 103 

Of Professor Alex. S. Herschel's contribution of Appendix III., . 104 

Appendix I. Separate Tables of observations of each of Twenty Gases ; pages 105 to 141. 
Appendix II. Tables of Gaseous Impurities, their characteristics and eliminations ; pages 142 to 149. 
Appendix III. Professor Alex. S. Herschel's observations under higher dispersions ; pages 150 to 159. 

Plate XL Elemental Gases, first as observed, and then as virtually freed from impurities. (Low 
Dispersion). 

Plate XII. Compound Gases, as observed, and with reference to their dissociations, as well as to unavoid- 
able practical impurities. (Low Dispersion). 

GENERAL INTRODUCTION. 

Of all the various spectra which the progress of experimental science has 
enabled man to observe in the present day, none are so rich, varied, and 
important, as those of gases. And no wonder ! for it is only when matter 
has been reduced to the gaseous condition, that it is able to specialise itself and 
write its character with much of its history in any otherwise smooth, undefined, 
continuous spectrum ; while, if in former times, men would have found it an 
impossibility to drive many of the more refractory substances into the state of 
incandescent vapour, what is there now anywhere on the surface of this earth 
which, in small quantity, can resist the action of a powerful and condensed 
induction spark of electricity ; and what application of that spark is so neat, 
elegant, convenient, and economical, as when it is employed in conjunction with 
so-called gas -vacuum tubes. 

In these tubes the infinitesimally small weight of the inch or two of almost 

VOL. XXX. PART I. p 



94 PROFESSOR PIAZZI SMYTH ON 

utterly rarefied gas which they contain, offers immense facilities to the electricity 
for dealing with it, so that a moderate size of galvanic battery, and a very little 
coil or sparking apparatus is all the observer needs to produce whatever light 
and heat he requires ; while a single small box with a dozen or two of thin 
glass vacuum tubes, each charged with a residuum of some particular gas, 
will enable him to inquire at any moment that he pleases, into the physical 
constitution of what makes up near half the universe. And this, too, without 
having to go through any chemical processes for procuring each gas whenever 
he wants it, in its extremest purity, and utter deadliness too, it may be. 

The heroic maker of the tubes ran that danger, and the subsequent fortunate 
possessor of them when made and hermetically sealed, has only to observe the 
spectra which the gaseous traces give out from the depths of their transparent 
prison-house, according to their labels, if duly attested and warranted by the 
maker, when the spark is passed through them. 

But what security, do you ask, can a mere label, though warranted by any 
maker, or even your own observed spectrum for that matter, give as to the reality 
and purity of the particular element of chemical matter supposed to be under 
examination? and do these tubes last? and do the gases in them never weaken, 
or change, or leak out, you wish to know ? Well, all that is really very 
important, both to be inquired into and to be published upon ; and it is, in fact, 
precisely what I have been looking into practically for a considerable length 
of time past ; with great hopes too at last of helping this mode of research 
to become, if not easier and more elegant than it has already been made by 
others, yet safer, truer, and more powerful than ever. 

The beginning of these latter-day attempts of mine was made in this way: — 
Twenty vacuum tubes of different gases and one or two volatilizable liquids and 
solids, such as alcohol, iodine and sulphur, were procured in duplicate from the 
late M. Geissler, in the form finally arranged with a capillary-central tube by 
the late Professor Plucker of Bonn. But when their spectra were found by 
me, generally faint, vague, and uncertain, a new arrangement and principle of 
viewing was invented, and twenty other pairs were procured on that different 
arrangement from M. Salleron in Paris. That new arrangement was founded 
on and constructed agreeably with the end-on principle of viewing, which I had 
the honour of setting forth before the Royal Scottish Society of Arts, in February 
1879, but which turns out to have been invented by Dr Von Monckhoven 
of Gand, in Belgium, several years earlier. Since then, a slight, but still further 
improvement has been made in my tubes, by giving them longer internal polar 
wires, to assist the electricity in traversing the necessarily large bulbs where its 
light is not wanted, and then throw itself with all its energy and along with any 
molecules of the gas it has caught hold of, into the capillary tube, and hurry 
along that with lightning like speed, and light as well. 



GASEOUS SPECTRA IN VACUUM TUBES. 95 

This capillary, when thus occupied by the incandescent rushing molecules 
and viewed end-on, presents a little disc, smaller than a pin's head, of light as 
to size ; but of exceeding brightness as to anything ordinarily seen in vacuum 
tubes. So bright, indeed, that when viewed under small spectroscopic dis- 
persion, one's eyes quail before the red and blue hydrogen lines as though they 
were glancing from the sun itself; while carbon bands appear more as solid 
things than haze ; and Nitrogen is simply a many coloured glory to behold. 

Of course that is a symptom in spectroscopy that those subjects will bear 
more dispersion : in which case by all means let them have it ; for only in that 
way can we ascertain the degree of importance of gaseous spectra. From that 
mere name of gaseous you might almost justifiably expect, that if there is 
anything sharp to be seen in them with a low power, it must of course become 
hazy and foggy with a high one, when made thereby to subtend a larger angle ; 
just as the edge of a cumulus cloud on the horizon, however well defined there, 
disappears as an edge, in soft formless vapour when we come close to it. 

But it is not so here in spectroscoping the ten thousandth part of a grain 
of electric-illuminated, rarefied gas. Take the Cyanogen pin's head of white 
light as an example ; stretch that little speck horizontally by spectroscopic 
power, just say to a finger's-breadth ; or, as it can be made to appear in 
angular space in even the smallest spectroscope, to half a degree in length ; 
and Ave have, with a broad slit, not much more than a very pretty spectrum, 
red at one end, citron in the middle, and violet at the other end ; with some 
hazy transverse bars of greater or less than the general brightness. 

Stretch it then a little more, say to 3 degrees ; and behold, by means of 
that, only in so far, increased scale of length from red to violet we now behold the 
alternations of more or less brightness, as seen before, explain themselves as a 
beautiful set of bands ; sharp as knife edges on one side, if the slit be rather 
narrowed and the focus improved, but indefinitely shading away at the other 
side ; whilst here and there are single lines burning and shining like linear 
suns ; only that in place of their being, in colour, all of them like our sun, 
yellowish-white, one is red, another orange, or citron, or green, or glaucous, 
or blue, or violet, or lavender, harmoniously with its spectrum place. 

So stretch the little pin's head of light more still, say to a length of 12 degrees. 
Why the bands are still more beautiful than ever ; still so sharp and solid on 
one side, but resolving themselves now into close linelets and ranks of the most 
needle-like lines on the other ; lines defying the powers of the micrometer to 
count their number, or equal them in thinness, or to separate them fully and 
clearly one from another. 

Wherefore now spare nothing ; stretch the luminous pin's head by prism 
power and magnifying power combined, until it forms an environing circle all 
round the observer, or subtends to him an angle of 360° ; and have you now 



90 PROFESSOR PIAZZI SMYTH ON 

destroyed the gift of rarefied gas in the spectroscope to look hard, solid, and 
sharp ? You have rather increased it, if your electric light keeps up well, and if 
your prisms are good, your spectroscope's slit perfect, and your objectives 
faultless ; for as you now survey the whole angular extent with the telescope 
of the spectroscopic apparatus, all appearance of dull smooth undefined haze is 
gone ; and almost everywhere, from the red on one side, right round the 
whole horizon until you come to the violet, — you see only luminous lines sharp 
and hard ; lines that will suffer no more splitting up under prism power ; no 
more dulling of their light by dispersion, though they may by the absorption of 
many more glass prisms. They have arrived then at their ultimate condition, 
and behold how exquisite they are ; how beautifully ordered in their ranks, 
how varied in their groups, how perfect in structure, contrasted in intensity, 
and indexed ready to one's hand by colour. It is almost endless work 
merely to admire them ; quite endless for poor, finite, human nature to try 
to measure them all. 

The only thing to compare with it would be a ring of the whole Heavens 
at once, on a bright and starry night when the firmament is shining with the 
countless glories of distant moving suns, the so-called fixed stars. And each 
of these innumerable bright lines in the electric-lighted spectrum of the 
cyanogen tube is as fixed, practically for ever, in spectral distance, the one from 
another, as each of those starry orbs ; so that a practical observer who should 
employ thereon throughout half his life all the angular forces of a mural circle, 
would by such observations be donating posterity with an heirloom of absolute 
knowledge of the most important and lasting kind. 

But even then the spectrum task is only at its beginning; for on each occasion 
that the spectroscopist changes a tube of one, for any other, gas, instantly the 
whole angular round is peopled with a new set of spectral but eternal lines ; 
each of which knows its own place, flies into it in less than the twinkling of an 
eye, and a new spectroscopic universe in lines of light is the result. Who would 
not, if he could, be an observer, to some extent, of such phenomena ; and, as 
for the cut bono use of it, if " the trains of thought " it leads the intelligent 
mind to, be not enough reward for ever, — then scientific history shows that 
the discovery of such a mine of accuracy in measure of place, and perfection in 
number, will be sure to have its practical applications in human education, and 
many industrial pursuits as well, before long. 

In this present paper permit me to begin simply as follows — 

PRACTICAL COMMENCEMENT DESCRIBED. 

This consisted, in my case, in the examination of the 20 tubes alluded to, 
with merely a single and simple prism power in the spectroscope, combined 



GASEOUS SPECTRA IN VACUUM TUBES. 97 

with a magnifying power of 10 on the telescope. So low a prism power was 
chosen at starting, to be certain of including everything even of the faint 
terminations, as well as of the brighter, but often discontinuous, portions of the 
middle, of any spectrum ; and it is far too small to allow of accurate measures 
of place. But as spectroscopic place, with some slight reference to colour and 
brightness, is the great identifier of spectrum lines, I have attempted to 
measure the place of everything seen in the tubes (at least with care), 
and have reduced mere instrumental numbers to the absolute scale of wave- 
lengths of light. These wave-lengths are however for practical convenience 
given as the numbers of such undulations in an inch British ; and go on 
increasing from 30,000 at the red, to 65,000 at the violet, end of the spectrum ; 
between which limits all our eye-observable phenomena will be found included, 
though photography takes account of many more. 

Examination of the observed quantities and elimination of" impurities" 

(Tables of 20 gas-vacuum tubes, spectroscopically examined. For these 
see Appendix I.) 

After the reader has examined each of the 20 gaseous spectra, as set forth in 
the Appendix above alluded to, the question may very properly be asked, how 
have they come out as to previous expectation, and to the promises made to 
inquire into the sufficiency of the makers' labels, for positive information 
touching the physical and chemical contents of each of the tubes \ 

The answer is, unfortunately, that a very little contrasting of one tabular 
statement with another will show superficial, perhaps, but most unexpected 
and annoying contradictions. We may have admired in one spectrum a 
certain group of lines, and thought them characteristic of the particular gas said 
to be contained there, especially as that was an elemental gas ; but these tables 
show the very same set of lines in another tube, and another, and another still, 
no matter what their labels may declare for them, and whether they refer to 
elemental or compound gases. 

This however implies a difficulty already well known in spectroscopy : — viz. 
that the chemists have not been able to prepare their products in sufficient 
purity, to resist the tests of the spectroscopist. But as they have done as well 
as they can, we, the observers, must now endeavour to eliminate the effects 
of practically unavoidable impurities by some virtual process ; and the first 
steps in this proceeding are fortunately very easy. 

Here for instance are four bright lines of Hydrogen, as measured by the 
late M. Angstrom. 

Red with place at 38,707, 
Glaucous „ 52,255, 
Violet „ 58,525, and 

Lavender „ 61,932. 



88 PROFESSOR FIAZZI SMYTH ON 

If these lines, viz. sharp lines at these places (the two first of them, and 
sometimes the third also, being usually very bright), are found in the tube of 
any other elemental gas, — you may mark them down as hydrogen impurities 
at once ; or if in the tube of a compound gas containing hydrogen in its com- 
position, it may be either an impurity, or the result of the dissociation of the 
compound by the electric spark, when each ultimate element gives out the lines 
belonging to it alone, as though the others were not present. 

Similarly there is a list of 4 Oxygen lines, given by Dr Schuster, — 

Orange at - . . . 41,254, 

Citron at . . . 46,729, 

Green at ... 47,659, 

Violet at . 58,156, 

and if only the 1st, 2d or 3d are found in any tube which is not oxygen itself, 
or has not oxygen as one of the chemically required constituents of its contents, 
— mark it off as an oxygen impurity there. 

But when we come to the third most important gas in all terrestrial nature, 
Nitrogen, there is a difficulty ; for Nitrogen at low electric temperature and 
seen under low dispersion, has no lines ; only bands ; and so numerous ! With 
very low dispersion they number 50 or 60 ; and at somewhat higher dispersion 
not less than 170 ! Moreover there is the further mental or moral or social 
difficulty that one-half of the greater spectroscopists of the age follow MM. 
Angstrom and Thalen in declaring that the above spectrum of bands seen in 
a nitrogen tube is not the spectrum of nitrogen at all, but of a compound, viz. 
Oxide of Nitrogen. 

Pare nitrogen gas they say has only one spectrum, and that is totally 
different to the above banded affair ; being a spectrum of a few sharp, piercingly 
bright lines, but which require a very powerful and condensed spark to enable 
them to show at all. When ordinary small sparks are employed, the nitrogen, 
they insist, falls into combination with oxygen, and exhibits bands, as usual with 
all oxides ; while oxygen is always present on such occasions, in consequence 
of the electric spark, however weak, dissociating the hydrogen and oxygen 
constituting the water of that innnitesimally small amount of moisture, which, 
it is averred, can never be perfectly driven out of the interior of glass tubes. 

Yet other equally great authorities follow the late lamented Professor 
Pujcker, and declare that the spectrum of 170 bands really is the Spectrum 
of Nitrogen, but at low spark-temperature ; and that most gases have two or 
more perfectly different spectra according to temperature. 

After trying both hypotheses on my tube observations, I incline to the 
latter of them ; . not so much from having been able to prove its absolute and 
perfect truth, as from having disproved the opposite view. Thus, in a Cyanogen 
tube, where there was no hydrogen line visible, there could not have been 



GASEOUS SPECTRA IN VACUUM TUBES. 99. 

any oxygen either, if that had to be derived, simultaneously with the hydrogen, 
from the decomposition of water ; and yet the bands of Nitrogen (said by the 
opposite school to be bands of an Oxide merely, because they were bands and 
not lines) were magnificently developed, broad, spreading, and true bands. 

So also with the Carbon bands of the same tube, derived from its Cyanogen 
combination of Nitrogen and Carbon being dissociated. The Carbon line 
spectrum consists of only eleven lines, and never shows except in a very 
powerful and condensed spark. But its band spectrum can be called up by 
any, even the smallest, spark ; and that band spectrum (said by Angstrom and 
Thalen to be necessarily belonging to an oxide of carbon because it is in 
bands) was brilliantly visible in this tube, where there could have been no 
oxygen for the carbon to oxidise with, i.e., if, as before, the oxygen had to be 
derived from the decomposition of water ; and the absence of hydrogen lines, 
inherently far brighter than those of oxygen, proved that such decomposition 
had not taken place. Hence, after virtually clearing my observations from 
Hydrogen and Oxygen, I proceeded in the same manner to get them free from 
traces of Nitrogen, Carbon, and the peculiar compound Carbo-hydrogen, — 
wherever these gases had no right of intended standing place. But still there 
were many lines left, and some of them very pronounced, common to several 
tubes with most diverse labels. What lines could they be ? 

By far the greater number turned out to be low temperature lines of 
Hydrogen. Almost a new class of lines in the spectroscopic world ; even 
denied by some persons, yet clearly visible simultaneously with the four great 
and almost classical Hydrogen lines ; which are properly high temperature 
emanations, but of such an intensity of vital force, as to be capable of living 
on, down through low temperatures also. And whenever they, the high- 
temperature lines, appeared in my low-temperature, but brilliantly lighted 
end-on tubes, there and then, in nearly the same proportions of relative 
intensity, appeared the crowds of the new low-temperature lines ; not three 
or four only, but rather three or four hundred. 

This discovery is involuntarily, but exemplarily given in a few of its principal 
features in the Tables of 

Appendix II. (see the end of this paper) ; 

which Tables show likewise the degree and manner in which impurities are 
distributed among the several tubes ; an instructive thing in itself. 

I have also prepared, but refrain here, for the cost's sake, from printing the 
practical deductions from Appendix I., in the shape of a set of resulting 
standard Tables of the places of leading features of gaseous spectra. The 
foundation for these places is always taken from the admirable observations of 



100 PROFESSOR PIAZZ1 SMYTH ON 

M. Thalen of Upsala, so far as they go. Then come some of my new lines 
from Appendix II., based in part on M. Thalen, and in part upon the old 
standard places of several well-known chemical flames, and a few Solar lines ; 
all of which are appended, for criticism and correction. 

This too is probably very necessary, where extreme accuracy is concerned, 
even among the oldest and longest known lines ; as particularly visible in the 
over large numbers for the place of the grand double line of Chlorine, in the 
best known Tabular statements thereof. For no such change of place, we hold, 
could have occurred in consequence of any varied mode of preparing the 
Chlorine, or treating it after being made ; but solely from error in reading oft* 
a micrometer screw, or printing the numbers from MSS. 

Changes with Time and Use. 

So much for endeavouring merely to secure correct numerical accounts at 
the instant, whether in support of, or opposition to, the makers' labels on the 
tubes. But now we must take up the second part of our promised answer, 
and testify somewhat as to the lasting power of the tubes, and also as to any 
physical changes occurring in them, after their contents had once been formally 
recorded and the record preserved. 

As to general lasting power of the tubes themselves, against the action of 
all ordinary and fair electric currents transmitted through them, and inclusive 
of an immense amount of sometimes not the gentlest handling in transferring 
them from their packing boxes to the electric holder, and vice versa, including 
too, several journeys by rail, — not more than one tube in twenty has failed, 
broken, or become dead — i.e., in all the specimens I have had longest, but 
whose glass-material was rather too soft, too easily fusible, and pervaded with 
some needless impurities. Very recently M. Salleron has adopted a harder 
glass ; chiefly for the sake of greater purity in the interior ; and that harder 
and less fusible glass is necessarily more brittle. But although it has given 
him an immense amount of trouble in the first formation of the tubes, — yet of 
six completed ones sent to me three months ago, they have stood all the trials 
well, and are exquisitely clear and transparent. 

Next as to the lasting power of the gaseous contents of the tubes, and their 
continued ability to keep on giving out the same spectrum under similar 
illumination, — the principal features of most of the tubes are undoubtedly 
maintained to a great, if not quite an absolute, degree ; and large changes 
have only occurred to two or three. But these have been note-worthy. 

To begin with the Cyanogen tube. It was first noted that one of the bulbs- 
was very prone to heat when in use ; then that the capillary's light, at first 
brilliantly white, had become faint and pink ; then that the bulbs were becom- 



GASEOUS SPECTRA IN VACUUM TUBES. 101 

ing fogged with brown colouring matter deposited on the inside, and finally the 
spectroscope showed that the carbon bands in its spectrum were disapiDearing, 
and various unknown and isolated lines were appearing instead, together with 
a growth or increase of Hydrogen manifestations. 

Some of the new lines could be made to disappear momentarily by 
introducing a Leyclen jar into the circuit, and were supposed therefore to 
belong to the compound gas Cyanogen ; but others could not be made so to 
disappear, and they proved to be the low temperature Hydrogen lines. Again, 
under special management of the condensed spark, the tube would for a short 
time blaze up vividly, and exquisite lines were then seen, thinner, sharper, and 
brighter than anything previous, — and they, from their places and relative 
intensities, must have been a part of the excelsior line-spectrum of Nitrogen. 

Several of these changes are noted in Appendix I. ; where two separate 
tables refer really to one and the same Cyanogen tube, but with a consider- 
able interval of time between them ; and another refers to a second tube of 
Cyanogen furnished to me by the maker on the same occasion as the first, but 
differing thus curiously in its spectrum ; viz., that while the first, in its earlier 
days, showed Nitrogen bands preponderating over those of Carbon, the second 
showed Carbon bands preponderating over Nitrogen ; but both of them were 
remarkable then for little or no hydrogen indications. 

The next tube to heat up, to change its light from white to pink, and to 
alter its spectrum, was Hydro-chloric acid. It had begun with chlorine lines 
brilliantly, some Hydrogen lines and faint Carbon bands. These last are now 
gone completely ; also, or even more signally, every one of the chlorine lines 
absolutely ; but the Hydrogen lines are all increased, and to such a degree as 
to compete with a pure Hydrogen tube for showing the Lavender as well as 
the other three principal and high temperature lines of Hydrogen, besides 
crowds of the new low temperature lines of that element. In fact I cannot 
distinguish its spectrum now from one of pure Hydrogen supplied by the 
maker as such ; but call its tube, for the sake of distinction without a difference, 
" the tube of artificial hydrogen." 

The pure chlorine tube still shows its chlorine lines, but they are becoming 
fainter ; and carbon bands and hydrogen lines have appeared, making its 
spectrum look very like what that of the Hydro-chloric acid was at first. 

Another tube that heats unduly, as if inclined for a change, and has much 
deposited haze in its interior, is Iodine : but no perceptible alteration of 
spectrum has yet been noted ; and because, perhaps, the maker put so large a 
quantity of solid iodine inside, that there is no chance of its all being dissociated, 
or converted into something else by my weak, small sparks, within any "moderate 
length of time ; — if Iodine is really, as some persons are beginning to suspect, 
not the elementary body which the chemists believe, but a compound. 

vol. xxx. PART I. Q 



102 PROFESSOR PIAZZI SMYTH ON 

Thus far then the few violent cases of change have shown a tendency in a 
feebly connected compound like cyanogen, and a doubtful element like chlorine, 
either to turn into hydrogen, or to develope so much of that brilliantly lighting 
gas, as to extinguish the fainter traces of anything else which may be left out- 
standing when chlorine dies, and hydrogen appears. But there is a case of far 
more ultimate importance, though much slower in working out, connected with 
Nitrogen ; and thus — 

I had observed with the Nitrogenous Cyanogen tube in its earlier days, 
that the Nitrogen bands there were clearer, more regular, even more Nitrogen- 
like than in the so-called pure Nitrogen tube itself ; but failed then to discover 
why ! Now, however, after comparing new and old tubes, the reason is perfectly 
plain. It was because at that time there was no Hydrogen in that tube; but in 
proportion as that and other tubes have been used, so they have developed 
Hydrogen; and though the widely separated 4 classical lines of Hydrogen may 
be eliminated easily, — the enormous numbers of the new low-temperature lines 
of that gas between red and blue are not so to be dealt with ; and they do in a 
manner take possession of, and tyrannise over, every band spectrum, utterly 
hiding or breaking up those fainter manifestations. 

With a new tube of Nitrogen, in the hard glass, there is a minimum of 
Hydrogen ; and the bands, as well as the groupings of bands, proper to Nitrogen 
throughout the red, orange, yellow and citron are the most delicate and 
beautiful series of gauzy veils, with sharp beginnings, imaginable, if viewed with 
a dispersion of 11° A to H, Mag. power 10 ; and for this one powerful reason 
specially, that " nothing interferes with them." But in an old tube of 
Nitrogen, though the same groupings of bands are seen beginning near the 
red hydrogen line, yet a little beyond that in the orange and yellow, the low- 
temperature lines of Hydrogen come in like a thicket ; and then no more 
Nitrogen bands are identifiable, until we get beyond low-temperature Hydro- 
gen's chief manifestations of its progeny, viz., into the blue and violet. 

So far as these two just described tubes may be trusted, time and use with 
the spark, would seem to have actually developed hydrogen in the older of 
them, either out of the glass matter of the tubes, or from the " occluded " 
stores of gas in the Polar wires, or more probably out of the Nitrogen gas 
contents ; and in that case, either by transforming Nitrogen positively into 
Hydrogen, or by dissociating it into its ultra-elements, of which the chief one 
must be Hydrogen, and the other something not yet recognised. These two 
latter hypotheses are of course dead against chemical theory as it now stands, 
but agree remarkably with some very different and more elevated lines of both 
spectroscopic and chemical research set forth a year ago by Mr Norman 
Lockyer to the Koyal Society, London. 

There would also appear to be an astronomical application, which, if not fully 



GASEOUS SPECTRA IN VACUUM TUBES. 103 

made before by some one else, opens up now some most noteworthy views in 
the quasi-vital chronology of the stars of heaven itself. 

Thus our Sun has been roasting for long geological, as well as human- 
historic periods in a temperature still higher than ordinary electric sparks ; 
and what do we find there touching these two critical gases, Nitrogen and 
Hydrogen ? That there is no Nitrogen, but overwhelming Hydrogen, in the 
Sun : or we might say, that its once supply of Nitrogen has been long since 
converted by continued supernal electric heating into Hydrogen. 

But in that case the beginning of the Sun's luminous history was probably 
marked by Nitrogen preponderating over Hydrogen ; and what do we find on 
recurring to Dr Huggins' remarkable observation on those agglomerating 
materials for Suns about to be, viz., the nebulae ? 

The answer is, " one faint hydrogen, but a much stronger and double 
nitrogen, line." 

On Recent Observations in Belgium. 

(Paragraph added during printing.) 

If the question be next put, " why only one line of each of those gases was seen, -when their 
usually admitted spectra contain several, or many," — the answer was not only given by Dr Huggins 
himself, to the effect that the visible line in the Nebula, was in each case the brightest of the several 
lines in the terrestrial spectrum of either gas ; — but special observations for the verification of, and with 
the effect most certainly of verifying, that great master-spectroscopist's view, have lately been made at 
the newly re-organized Royal Observatory of Brussels. 

M. Fibvez, the spectroscopic astronome-adjoint there, had already communicated several researches 
on allied points in spectroscopy to the Academie Royale de Belgique, when he took up this question, 
with results now published in the Academy's Bidletins, 2 me serie, torn. lxix. N° 2, 1880; and his 
apparatus was so vastly superior to mine, as to supply some much desired data for its possible future 
extension and improvement. Thus, while M. Fievez employed end-on tubes very like my own, he 
illuminated them, not by such wretched little sparks as I was confined to by private economy, viz. 
sparks generally under 1 inch, or even half an inch long, — but by sparks 20 inches long, procured from 
a very large induction coiL excited by a Bichromate battery of 8 couples (size not mentioned); and 
these sparks occasionally intensified by the use of a condenser of 6 square yards of surface, employed 
sometimes in tension and sometimes in quantity. 

Now as my condenser consisted only of one quart-sized bottle, and I was even afraid of using that 
much lest the glass tubes should crack, — I wrote to M. Fievez asking how he contrived to ensure the 
safety of his tubes, when tried in such almost fearful light and heat. 

He kindly replied " that he always began by very slowly immersing the zincs of the battery into 
the acid solution, producing only a feeble current." Some instants afterwards he introduced the con- 
denser in tension into the circuit, and then immersed the zincs a little more. Lastly he disposed the 
condenser in quantity (as a single element) ; but he was careful to keep it acting in that manner only 
for a few minutes because the heating of the capillary if the so illuminated vacuum tube became too 
considerable. He further added, that a tube of Hydrogen-vacuum which had served for many experi- 
ments of that kind, presents now a deposit of metallic aluminium (derived from the electrode wires) at 
one of its extremities. 

Of course the brightness of the spectra presented by M. Fievez' tubes under 20 inch sparks or their 
condensed equivalents, was magnificent, delightful to the observer to behold, and greatly promotive of 
exactness in any mensurational applications. Of course also his Nitrogen tubes showed the sharp linear, 
not the faint band, spectrum of that gas ; and equally of course the 4 classical, high-temperature lines of 



104 PROFESSOR PIAZZI SMYTH ON 

Hydrogen are the only ones he mentions seeing in the Hydrogen tuhe ; while amongst them the palm of 
brilliancy is is not with the red, as it is so often in small sparks, hut in the more refrangible region of 
the glaucous Hydrogen line at 52, 255 W.N. Place. 

M . Fievez' observations then were conducted on an electric stage quite above that on which I worked ; 
and he shows how any one else may attain to the same. I will therefore only add, that I believe there is 
another stage below mine again, which would yield most important results for some of the physics of the 
faint Comet ary, and Sidereal systems, could it he practically realised and well worked; witness the fol- 
lowing very recent case for it. 

M. Jamin lately showed, in the Academy of Sciences in Paris, that the origination of the "proper" 
light in a Comet's tail, must be the illumination of its carburetted constituent molecules by electric dis- 
charges of some kind, mainly because the only other known possible method of illumination, viz. by 
combustion, was absurd and utterly inapplicable under the circumstances. Professor Young, of Prince- 
ton, U.S., the present Astronomer Royal at Greenwich, and others have on the contrary proved, by 
observation, and spectroscopic measurement of place, that the carburetted spectrum exhibited by a 
Comet's tail is not that of the carbon-band order of electric illumined gas-vacuum tubes, — but is that of 
combustion of coal-gas and common ah in the blue base of any ordinary burning flame. 

Now, as mentioned in the following pages, I have already found, on merely shortening and thicken- 
ing the wire forming the outer helix of my very moderate induction coil, (and thereby reducing the 
intensity of its sparks) that the brighter features of certain carbo-hydrogen combustion bands could be seen 
in an olefiant-gas vacuum tube, and less than before of the carbon band electric-lighted tube spectrum as 
usually known. Could we then, — by employing some very different method to the induction coil, of 
producing luminous electricity, as by the friction machine, Holtz's machine, or others, — so much further 
still reduce the intensity, while still keeping up the quantity, of the illuminating spark, as to render 
visible to us the combustion spectrum only, without any trace of the only hitherto known tube spectrum 
(which is the electric carbon-band spectrum) of a carbo-hydrogen gas — we should accomplish this ; viz., 
we should have reached the chief physical conditions of visibility of such a Comet as Tebbdtt's great 
Comet of 1881, and harmonised at the same time the present apparently utter oppositions of M. Jamin's 
theory versus Professor Young's and Mr Christie's observations. 

Of Prof. Alex. S. Herschel's Contribution of Appendix III. 

With the same spectroscope, tubes, and sparking apparatus employed by my- 
self, but with more powerful prisms inserted, and also some new tubes of his 
own, many observations have been made from time to time con amove by my 
friend Prof. Alex. S. Herschel ; and when I found that he had very original 
theoretical ideas as to the arrangement of the lines and bands in many spectra, 
I invited him to lose no time in communicating them to the Royal Society, 
Edinburgh. This he kindly promised to do, if agreeable to the Society ; and 
although several other modes of presenting his views occurred to us, and were 
discussed, he preferred the method of contributing an Appendix to the present 
paper ; on the clear understanding, however, that he is not necessarily bound 
by anything which I have written in the preceding part of this, or in any other, 
spectroscopic paper, but by his own portion only, viz., Appendix III. 



GASEOUS SPECTRA IN VACUUM TUBES. 



105 



APPENDIX I. 



SEPAKATE TABLES OF OBSERVATIONS OF EACH OF 20 GASES. 

Series of 20 End-on Tubes, observed in 1879 and 1880 with Aurora spectroscope, small dispersion 
(1 prism of 52° refracting angle of white flint, having 3°'3 Disp. from A to H), and small intensity of 
sparks (generally under 1 inch long, and latterly under # 3 inch, but from an ordinary so-called 
2-inch spark induction coil purposely reduced in intensity by replacing its outer helix of long thin 
wire, with another of thicker wire and less length ; bichromate battery of 5 pots, the zincs measuring 
2"-5x4" each). 



Names of the assumed 
Tube Fillings. 


Symbol. 


Elemental. 


Compound 


Names of the Compounded 
Elements. 


Air, 




N + + &C. 




Mixture j 


Nitrogen, Oxygen, Watery 
Vapour, &c. 


Alcohol, . 




C.H.O 




Compound 


Carbon, Hydrogen, and Oxygen 


Ammonia, 




NH, 




Compound 


Nitrogen and Hydrogen 


Carbonic Acid, 




co 2 




Compound 


Carbonic Oxide and more 
Oxygen 


Carbonic Oxide, 




CO 




Compound 


Carbon and Oxygen 


Chlorine, 




CI 


Elemental 






Cyanogen (old),' 




CN 




Compound 


Carbon and Nitrogen 


Cyanogen (very old), 


CN 




Compound 


Carbon and Nitrogen 


Cyanogen (second example), 


CN 




Compound 


Carbon and Nitrogen 


Hydrochloric Acid, 




HC1 




Compound 


Chlorine and Hydrogen 


Hydrogen, 






H 


Elemental 






Iodine, 






I 


Elemental 






Marsh Gas, 






CH 4 




Compound 


Carbon and Hydrogen 


Nitrogen, 






N 


Elemental 


... 


... 


Nitrous Oxide, 






N 2 




Compound 


Nitrogen and Oxygen 


Olefiant Gas, 






C. 2 H 4 




Compound 


Carbon and Hydrogen 


Oxygen, . 









Elemental 






Ozone, 






o 3 


Elemental 


Allotropic 




Salt-water, 






H 2 + Na 




Compound and Mixture 


Hydrogen, Oxygen, and Salt 


Water, . 






H 2 




Compound 


Hydrogen and Oxygen 


vol. x: 


£X. 


PART I. 








R 



106 



PROFESSOR PIAZZI SMYTH ON 



AIR. End-on Gas Vacuum Tube. Observed Sept. 26 and 27, 1879. 
A Mixture of N + 4- Watery Vapour, &c. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed 
for Air. 


Crimson- 
Red. 


Part 1. 

Faint line 

Very faint band, .... 
Faint band 


5 
0-3 

7 


Ml 
M 


33 790 

36 764 

37 078 
37 298 
37 725 


I 


Nitrogen 
Nitrogen 

Nitrogen 


None ; and no 
solar telluric 
lines appear. 


Red. 




Scarlet-Red. 


Band, 


ro 


Hj 

"I 

1 


37 936 

38 206 




Nitrogen 






Band, 

Red hydrogen, .... 


1-0 
5-0 


38 304 
38 715 
38 715 


Hydrogen 


Nitrogen 






Part 2. 

Red hydrogen, .... 

Band, 


5-0 
1-0 


I-', 1 


38 699 

38 803 

39 063 


Hydrogen 

1 

! 
! 
1 


Nitrogen 






Light-Red. 


Bright band, faint cleft down ) 
middle, \ 

Do. do. do. . 


1-5 
2-0 




39 219 
39 557 
39 687 
39 937 


Nitrogen 
Nitrogen 






Do. do. do. . 


2-0 


m 1 


40 093 
40 509 


Nitrogen 






Orange. 


Narrower band, .... 


2-0 


M| 

ih'l 


40 639 
40 842 


! 


Nitrogen 






Distinct line, with black line or 
space on either side, . 

Unequal double line, . . ] 


1-0 
1-0 
2-0 


40 967 

41 149 
41 279 


Oxygen ? 
Oxygen ? 
Oxygen 


? 

2 




Close band of lines, 


2-0 


mil | 


41 409 
41 462 


j 

[ Carbon 
\ Hydrogen 

! » 

\ Hydrogen 
? 

V Hydrogen 


Nitrogen 






Strong beginning of narrow band, . 
Strong line beginning a graduated ) 
band, . . . . > 
End of that band, . . ) 
Bright hazy line, .... 
Space intervenes full of lines, 
Strong ending line of above, . 
New band of many thin lines begins, 
Strong ending line to that band, . 
Thin line follows, .... 
Band ; begins faintly, . 

culminates in strength, 

ends faintly, 
A notable black line follows. 


1-5 

2-0 

1-0 

2-5 

1 

2-5 

1 

2 

0-5 

1 

3 

1 


| B ,| 

ill 

urn 

mi 
1 

!' ! 


41 707 

41 881 

42 115 
42 161 

42 511 
42 608 
42 815 

42 929 

43 039 
43 141 
43 288 


Nitrogen ? 

Nitrogen 
Nitrogen 

Nitrogen 


? 


Yellow. 




Part 3. 

Band begins after a black line, 
Strong bright line ends band, 


1 
2 


! A 1 


43 374 
43 561 


f 
\ 


Nitrogen 






Bright line perhaps double, . 


2-5 


1 


43 700 


I New 
| Hydrogen 








Band of close lines, 


1-5 


mi 


43 816 

44 017 


1 

Hydrogen 
Hydrogen 


Nitrogen 






Citron. 


Hazy line, ..... 
Thick bright line, .... 
After this a notable broad dark 


1-5 
3 


i 111 


44 152 
44 351 


Nitrogen ? 
Nitrogen ? 






space. 

1 Line 

Faint features in the dark 1 Line 

space, . . . y Line 

( Band 


0-5 
1 

0-5 
0-5 


i 

i 
i 

i 


44 523 
44 685 

44 911 

45 112 


I Hydrogen 
) Hydrogen 


Nitrogen 






A band begins a brighter green ( 
region, . . . . j 


2 


hJ 


45 245 
45 478 


I Carbon 


Nitrogen 






A fainter band, .... 


1-5 


i=i j 


45 584 
45 846 


1 

! 


Nitrogen 






Do. do 


1-5 


H| 


45 979 

46 234 


Nitrogen 





GASEOUS SPECTRA IN VACUUM TUBES. 

A IK. End-on Gas Vacuum Tube — continued. 



107 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed 
for Air. 




Part 3 — continued. 
















Hazy line, beginning of many such, 


1 


I|: 


46 358 




Nitrogen ? 


? 




Stronger hazy line, 


1-5 


in : 


46 551 




Nitrogen ? 


? 




Do. do 


1-2 




46 706 




Oxygen ? 


2 


Citron. 


Do. do 


T2 




46 878 




Nitrogen 






Do. do 

Do. do 


1-2 
1-0 


II! * 


47 006 
47 176 




Nitrogen 
Nitrogen 








Do. do 


1-0 


ill 


47 298 




Nitrogen 






Do. do 


1-0 


1 


47 420 




Nitrogen 






Thicker hazy line, 


2-0 


ill 


47 582 




Oxygen 






Dark space intervenes. 














Green. 


Thick hazy line, .... 


17 


ill 


47 845 


Hydrogen 


Nitrogen 






Band, 


1-5 
( 0-5 


i 


48 129 
48 312 

48 433 


i 


Nitrogen ? 


? 




Very faint hazy lines, 


) 0-5 
) 0-5 
( 0-5 


i 
i 
i 


48 554 
48 677 
48 759 


J 


Nitrogen 






Strong hazy line begins a band, 


3 


) ( 


48 900 


) 








Second line therein, 


2 


III::. 


49 041 


> Carbon 








Said band slowly fades out, . 
Line in region of faint haze, . 


0-5 
1-5 


i ( 


49 478 

50 121 


i 


Nitrogen 








Fainter do. do. 


0-7 


.:|, 


50 481 




Nitrogen 






Do. do. do. 


0-7 




50 668 


( Carbon ? 
\ Hydrogen 








Strong line begins a band, 


2-0 


| h. j 


51 071 




Nitro cr en 






Said band ends weakly, 


0-3 


51 297 


1 








Faint hazy line, .... 


1-0 


1 


51 526 


Hydrogen ? 








Another like it, . 


1-0 


ill 


51 710 


2 


Nitrogen ? 


? 




Glaucous hydrogen, 


4-5 


1 


52 243 


Hydrogen 






Glaucous. 


Part 4. 
















Beginning of a faint band, 


5 




52 594 


Carbon 








Strong beginning of another band, . 


1-5 




52 812 


1 


Nitrogen 






Tail of above band ends, 


0-3 




53 235 








Strong beginning of another band, . 
Tail of band ends here, . . 


2-0 
0-3 




53 877 

54 284 


I 


Nitrogen 






Hazy line, 


1-3 


HI /" 


54 507 


) 








Strong beginning of another band, . 


2-0 




54 678 




Nitrogen 






Tail of baud ends here, . 


0-3 




55 046 


) 








Strong beginning of another band, . 
Tail ends hereabouts, 


2-0 
0-3 




55 613 

56 178 


! 


Nitrogen 






Blue. 


Strong beginning of another band, . 


2-0 




56 707 


I 


Nitrogen 






Tail ends here, .... 


0-3 




57 243 
















( 


Nitrogen 






Another band begins strongly, 
Tail ends here 


1-7 




57 695 


\ 


(No Oxygen 
visible here 




Indigo. 


0-3 




58 211 
















( 


at 58 156.) 






Violet hydrogen 


2 


1 


58 537 


Hydrogen 






Part 5. 
















Strong beginning of band, 


1-5 




58 725 


| 


Nitrogen 




Violet. 


Tail of band, 


0-3 




59 005 








Beginning of another band, . 


1-5 




59 667 




Nitrogen 






Tail of that band 


0-3 




60 020 


1 






Faint hazy line 

Strong beginning of another band, . 


0-5 
1-0 




60 388 
60 625 


Carbon? 

j 


Nitrogen 








Tail thereof 


0-3 




61 026 






Strong but hazy beginning of 
















another band, .... 


0'9 




61 364 


1 


Nitrogen 




Lavender. 


Tail thereof, 


0-3 




61 811 






Hazy beginning of another band, . 


0-6 




62 092 




Nitrogen 






A hazy band, .... 


0-4 




62 754 




Nitrogen 






Very faint band, .... 


0-3 




63 734 




Nitrogen 






Doubtful band, .... 


o-i 




64 465 




Nitrogen 





In the Air-tube no change is perceptible over all the above Part 5 range (or known anywhere else) on changing the poles. 
No line conies in, such as M. de Boisbaudran's strong a line in his open-air spark near the Negative Pole. 

The impurities traced here are Hydrogen (strongly, though in few lines only) and Carbon (weakly) ; the constituents are 
Oxygen not very brightly ; but Nitrogen overpoweringly in the number and often the strength of its bands. 

Nothing of importance left for Air as a compound. No Telluric lines, as in the Solar spectrum, are indicated. 



108 PROFESSOR PIAZZI SMYTH ON 

ALCOHOL TUBE, End-on. 

C 2 H G 0. Carbon, Hydrogen, and Oxygen. 

Observed 12tb July 1879. From tbis date a more rigid system adopted of dividing tbe Spectrum into 
5 parts, with special references, settmgs, and focussings for eacb. Tbe references kept strictly apart from the 
observations. 



Colour. 


Subject of Observation. 


intensity. *%£■ 


1 

W.N. Place. 




Impurities. 


Constituents 
dissociated. 


Unclaimed 
features left 
for Alcohol. 
















None beyond 
















ordinary Carbo- 
















Hydrogen, and 
















a portion of 




Parts 1 and 2 of Spectrum — 












Marsh-gas. 


Scarlet- ( 


Faint line, .... 


1 | 


38 351 






Carbon 




Red. | 


Red hydrogen, 


8 I 


38 707 


X 




Hydrogen 




Light-Red. 


Broad haze band, . 


n ! 


38 780 

39 968 


! 


? 






' 


Narrow haze band, 


2 


40 141 






Carbon ? 






Hazy band, .... 


3 mi - 


40 268 
40 562 


j 




Carbon 






Do 


3 \m\ ■ 


40 669 
40 892 


! 




Hydrogen 






Do 


3 m 


40 954 

41 335 


1 




( Hydrogen 
j (no Oxygen 














( appears. ) 
















( Carbo- 




Orange. - 


Do 


4 \m\ 


41 384 
41 626 


!« 




) Hydrogen 
1 and 
( Hydrogen 


Carbo- 
Hydrogen 




Strong edge of band, 


5 &: j 


41 810 


X 


i 


Carbon 






Shaded-olf edge of same, 


42 068 








Hazy band, .... 




42 135 
42 315 


1" 




Hydrogen 






Hazy band, .... 


3 m 


42 467 

42 658 


I 




? 




. 


Excessively thin line, 


0-5 | 


42 725 




I 








Hazy band, .... 


2 mi ' 

4 1 


42 772 

43 047 


1 

)'■ 
5 




Hydrogen 






Yellow line (not sodium), 


43 157 




Hydrogen 






Part 3 of Spectrum — 






43 170 








Former yellow line, 


3 1 


43 183 








Yellow. - 




1 | 


43 589 






Hydrogen 






Strong yellow line, . 


4 1 


43 737 


X 




Hydrogen 






Haze intervenes, . 
















Marked line, .... 


2 X 1 


44 146 






Hydrogen 






Hazy line, .... 


3 ill 


44 364 






Hydrogen 




s 


Broad faint haze band, . 


1 -1 

I 1 


44 549 

45 007 


I 


Nitrogen ? 


+ Hydrogen 




f 


Line probably citron 1 of blow- 


45 079 






I Carbo- 


Carbo- 




pipe, ..... 






( Hydrogen 


Hydrogen 




Tube's citron band, sharp be- 
ginning, . . 


| 4 „ 


45 264 


X 


} 


Carbon 




Citron, -j 


Do. do. extends faintly to 


| 1 


45 383 




i 






Haze further intervenes, 
















Haze band 


2 m 


45 990 






I Carbo- 
( Hydrogen? 


Carbo- 
Hydrogen 


I 


Faint hazy region, . 


1"5 ;;; 


46 180 
46 867 




1 


i Carbo- 

j Hydrogen? 


Carbo- 
Hydrogen 




Sharp beginning of band, 


2 ::::. 


47 067 






Carbon 






Faint hazy region, . 


1 


47 540 
47 961 




I 


Hydrogen ? 






Faint hazy region, . 


1 ... 


48 122 
48 375 




1 


Hydrogen ? 






Green band's sharp beginning, 


6 ::::. 


48 792 


X 




Carbon 




Green. 


Blow-pipe's green giant therein, 
Part 4 of Spectrum — 


8 I 


49 173 


r 


49 176 


j Carbo- 
( Hydrogen 


Carbo- 
Hydrogen 




Green giant in tube, 


7 I 


49 179 


) x 




j Carbo- 


Carbo- 




Its second line do., 


3 1 


49 520 


X 




( Hydrogen 


Hydrogen 



GASEOUS SPECTRA IN VACUUM TUBES. 

ALCOHOL TUBE, End-on— continued. 



109 



Colour. 



Glaucous. - 



Subject of Observation. 



Blue. I 
Indigo. 

r 

Violet, -j 
I 



Lavender. 



Part 4 of Spectrum — continued. 
Haze extends thus far, . 

Faint band, 

Fainter haze band, 

Hazy line, .... 

Glaucous hydrogen, 
Blue band, tube's, 

Tube's blue band extends 

faintly to, . 
Line, supposed blow-pipe's blue 
band 1, . . . 

Do. do. line 2, 



Do. 

Hazy band, 



do. 



line 3, 



Hazy line, 

Band begins sharply, 

Part 5 of Spectrum — 

Violet band begins sharp, 
Faint violet band, . 

Violet hydrogen, 

Band, supposed blow-pipe's in 
tense, .... 

Line, supposed of Marsh-gas 

Another, 

Another, 

Hazy followers, from near to, 
Do. do. up to, 

Lavender band begins, . 

Do. do. ends, 
Haze band, 
All haze ends, 



Intensity. 


Appear- 
ance. 


0-5 




0'5 


-1 


0-3 


'J 


1 


6 

4 


■::. 


| 0-5 




!■ 


1 


1 


1 


0-5 


1 


1 




1 
3 


.. :'= 


3 
2 


:■- 


2 


1 


I 2 




2 

1 
1 


1 

1 

1 


| 0-3 


- | 


2 
0-5 

1 
0-2 


=f 



W.N. Place. 



50 443 
50 527 
50 848 

50 989 

51 322 

51 527 

52 240 

52 502 

53 172 
53 631 

53 856 

54 050 

54 244 

54 797 
56 330 



56 320 

57 830 

58 515 

58 895 

60 245 
60 529 
60 740 

60 970 

61 239 

61 592 

62 002 

62 570 

63 068 



Impurities. 



56 325 



Constituents 
dissociated. 



Carbon 
Carbon ? 

Carbon ? 
Hydrogen 

Hydrogen 
Carbon 

Carbon 

Carbo- 
Hydrogsn 

Carbo- 
Hydrogen 

Carbo- 
Hydrogen 

Carbo- 
Hydrogen 
Hydrogen 

Carbon 



Carbon 

Hydrogen 

Carbo- 
Hydrogen 
Marsh-gas 
Marsh-gas 
Marsh -gas 
Marsh- gas 
Marsh -gas 



Carbon 
Carbon 



Unclaimed 
features left 
for "fftcoholr 



Carbo- 
Hydrogen 

Carbo- 
Hydrogen 

Carbo- 
Hydrogen 

Carbo- 
Hydrogen 



N.B. — Olefiant-gas tried after the above, has perhaps the faintest trace of No. 1 of Marsh-gas at 60 245. But Marsh-gas 
has that whole group splendidly, brilliantly, viz., from 60 243 to 62 139 ; or shows more of the lines and bands of it by far 
than Alcohol ; and Alcohol again than Olefiant-gas. Yet Olefiant-gas is far nearer in chemical constitution to Marsh-gas 
than Alcohol ! 

The Carbon, Hydrogen, and Carbo-hydrogen constituents of this tube appear abundantly ; but none of its Oxygen con- 
stituent ! 

More remarkable, however, is it to note in this Alcohol tube, over and above the carbon bands, known positively to be 
electric-spark carbonic manifestations, the appearance of other bands of a carburetted kind, but only known, or acknowledged 
hitherto as combustion, not as electric, bands. Yet here they are undoubtedly in nothing but faint, weak, electric 
illumination. 

Thus the green band, whose sharp beginning is quoted above as being in W.N. Place = 48 792, is the electric, vacuum- 
tube, carbon or carbonic-oxide band of all electric observers ; and has never been seen in any lamp or candle burning freely 
in the open air. 

But the other green band, which begins with a strong line at 49 176 of W.N. Place, is the band which may be seen in the 
blue base of every carbo-hydrogen flame, lamp, candle, or anything else ; has been taken hitherto as an example and type of 
a combustion spectrum ; and yet is the green band which is seen in comets, where, on principle, there can be no combustion, 
but may be faint electrical currents. 



110 



PROFESSOR PIAZZI SMYTH ON 

AMMONIA. End-on Tube. NH 3 
July 28, 1879. 



Colour. 



Subject of Observation. 



Crimson- 
Red. 



Red. 



Scarlet-Red. 



Scarlet-Red. 



Li^ht-Red. 



Orange. 



Yellow. 



Yellow. 



Part 1. 

First faint red band, 
Faintest haze intervenes, 

Faint band, . 

Stronger and denned band, 

Band well defined, 

Part 2. 

Band defined at sides, . 

Red hydrogen line, brilliant, 

Band, .... 

Band split down middle, 

Do. do. 

Bright hazy line, . 
Do. do. 
Do. do. 

Band, .... 



Sharp line, 
Line in haze, 
Haze intervenes, 
Line in haze, 

Line, 

Haze intervenes, 

Line, 

Line, 

Haze intervenes, 

Line, 

Haze intervenes, 

Line, strong, 

Fainter line, . 
Still fainter line, 

Bundle of thin lines, 
Haze intervenes, . 
Bundle of thin lines, 

Thin line, 



A signal and strong yellow line, but 
tested not to be in Sodium's place, 
but slightly beyond it, 



Part 3. 

The former yellow line, 

Bundle of thin lines, 

Line, 

Line, 

Strongest line yet, 



Definition generally admirable. 



Intensity. 


Appear- 
ance. 


0-6 


= 


1 


-I 


1-7 


l-l I 


2 '5 


■1 


2-5 


m | 


8 


■ 


2-5 


H j 


2-5 


■ { 


2-5 


Ell j 


2 
2 


.. .. I!,!i 


2 


iilii 


2 


H ! 



W.N. 


Place. 


34 


010 


36 


858 


37 


202 


37 


422 


37 


762 


37 930 


38 


168 


38 272 


38 


558 


38 


707 


38 


802 


39 


055 


39 


222 


39 


498 


39 


628 


39 


950 


40 


132 


40 


309 


40 487 


40 


652 


40 795 



Impurities. 



3 


1 


1-5 


1 


0-5 




2 


iilii 


2 


I 


0-5 




2 


1 


2 


1 


0-5 




2 


1 


0-5 




3 


1 


1-5 


1 


1 


i 


2 
0*5 


mil 


2 


mil 


1 


i 


1' 


i 


2 

1-5 

1 


i 

llll 

1 


1-8 


1 


3-5 


1 



40 
41 

41 

41 

41 

41 



943 

128 

231 
386 
479 

670 



41 838 



42 114 



42 
42 

42 

42 
42 



186 
298 

497 

792 
934 



43 156 



43 
43 
43 
43 
43 



161 
286 
410 
532 



Constituents 
dissociated. 



Oxygen ? 



43 158 



Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 

Hydrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Hydrogen 
Hydrogen 
Hydrogen 

Nitrogen 



Hydrogen ? 
Hydrogen ? 



Hydrogen 
Hydrogen 

Nitrogen ? 

Hydrogen + 
+ Nitrogen ? 

Hydrogen + 
+ Nitrogen ? 

Nitrogen ? 

Nitrogen ? 

Nitrogen ? 

Nitrogen ? 
Hydrogen + 
+ Nitrogen ? 



Unclaimed for 
Ammonia. 



Nitrogen ? 
Nitrogen ? 
Hydrogen 
Hydrogen 



Hazy region intervenes, nearly resolvable into lines. 



GASEOUS SPECTRA IN VACUUM TUBES. 

AMMONIA. End-on Tube. NH 3 — continued. 



Ill 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed for 
Ammonia. 


Yellow. 


Part 3 — continued. 

Line on heavy black ground, . 
Haze intervenes, .... 
Line in haze, .... 
Line, ...... 

Broad line or bundle of thin ones, . 

Thinnest possible line, . 


1-1 
1-1 

1-5 
3 

0-2 


1 

i 


43 874 

43 991 

44 114 
44 241 
44 358 
44 451 




Nitrogen ? 

Nitrogen ? 

Hydrogen 
Hydrogen + 
+ Nitrogen ? 

Hydrogen 


? 


Citron. 




Line 


1-5 


1 


44 548 


I 


Hydrogen + 
+ Nitrogen ? 






Line, 

Very thin line, .... 
Line, ...... 


1-5 

0-3 
1-3 


1 

i 
1 


44 686 

44 798 
44 882 


1 


Hydrogen + 
+ Nitrogen ? 
Nitrogen '{ 
Nitrogen ? 






Haze band, 

Broad haze band, with line in 
centre, ..... 

Faint haze intervenes, . 


07 

( 0-5 

2 
( 0-5 

0-2 


M 


44 984 

45 288 
45 268 
45 384 
45 468 


1 
I 


Nitrogen 
Nitrogen 






Semi-resolvable haze, 


0-5 


- 1 

in 


45 572 
45 908 










Bundle of thin hazy lines, 


0-5 


46 016 










Haze band, ..... 


1-0 


a | 


46 098 
46 250 


j 


Nitrogen 






Faint haze band, .... 


0-5 


,| 


46 298 
46 412 


j 


Nitrogen 






Stronger haze band, 


1-0 


M j 


46 458 
46 602 


i 


Nitrogen 




Citron. 


Hazy line, 

Strong hazy line 

Bundle of thin lines, 

Line, ...... 

Hazy line, ..... 

Hazy line, ..... 

Bundle of lines, .... 

Single thin line, .... 

Bundle of thin lines, 


1-0 
1-5 
0-8 
1-0 
0-8 
0-5 
1-7 
1-0 
1-5 


:i: 

ill 

llll 
1 

" llll 
llll 


46 730 

46 864 

47 020 
47 158 
47 298 
47 442 
47 590 
47 730 
47 870 


Oxygen ? 
Oxygen ? 


Nitrogen ? 
Nitrogen ? 
Nitrogen 
Nitrogen ? 
Nitrogen ? 
Nitrogen ? 
Nitrogen 

Nitrogen 


2 


Green. 




Region of faint semi-resolvable haze, 


0-5 


. j 


47 980 

48 096 
48 160 


i 


Nitrogen 






Bundle of close thin lines, 


2 


iiiiii 


48 216 
48 280 
48 320 


s 


Nitrogen 






Faint hazy line, .... 
Line in fainter haze, 

Do. do 


1-0 
1-0 
0-8 


III 


48 428 
48 576 
48 628 




Nitrogen 
Nitrogen 
Nitrogen 






Faint haze band, .... 


0-4 


\| 


48 680 
48 788 


I 

$ 


Nitrogen 






Strongest line yet ; and not altered 
by condenser 


| 3-5 


1 


48 884 


Carbon ? + 


Nitrogen ? 


Ammonia ? 




Another strong line, with de- 
creasing haze beyond it, to 
further than the place of the 
Blow-pipe's Green Giant, and 
without symptom thereof, . 


- 3 
J 


i. 


49 020 




Nitrogen ? 


Ammonia ? 


Green. 


Part 4. July 31, 1879. 

End of semi-resolvable haze, . 
Thin line, ..... 
Strong line, beginning band of 
thin ones, ..... 
End of above band, 


0-3 
1-0 

| 2-0 

0-5 


i 
t. 

in 


49 910 

50 012 

50 170 
50 424 


! 


Nitrogen 
Nitrogen 




Glaucous. 



112 



PROFESSOR P1AZZI SMYTH ON 

AMMONIA. End-on Tube. NH,— continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- w N place 
ance. 


Impurities. 


Constituents 
dissociated. 


Unclaimed for 
Ammonia. 


Glaucous. 


Fart 4 — conti lined. 

Distinct line, .... 
Line, beginning haze band, . 
Strong line, ..... 
Faint line close up to above, . 


10 
2-0 
2-0 
1-0 


1 
Ph. 

1 

1 


50 512 

50 668 

51 064 
51 128 


Carbon ? 


Nitrogen 

Nitrogen ? 
Nitrogen ? 
Nitrogen ? 


Ammonia ? 

2 




Thin line, 

Next double line followed by extra 
black space, . 


1-0 
|,0 


'■1 


51 252 
51 472 
51 534 


! 


Hydrogen 


? 




Decreasing baud of thin faint lines, 


0-7 


!, | 


51 660 

52 030 


1 


Nitrogen 






Glaucous Hydrogen vividly bright, 


10 


I 


52 252 




Hydrogen 






Faint haze beyond, 
Line beginning band, 


0-3 
2 


'" III:, 


52 590 
52 782 


Carbon ? 


Nitrogen 






Liue beginning grand band, . 

End of that band at 

Faint line, ..... 


3 

0-5 

1-0 


Ph. 


53 841 

54 310 
54 480 


I 


Nitrogen 
Nitrogen ? 


? 


Glaucous. 


A notable wide and blue double 
line, ...... 

Line beginning graduated band, 
Line beginning grand blue band, . 

Part 5. August 2, 1879. 

Similar line, beginning a similar 
band, ..... 


11 

2-5 
3-0 

}• 

3 
| 0-3 


1 

1 

t 
1 


54 670 

55 012 

55 556 

56 620 

57 555 


1 


Nitrogen ? 

Nitrogen 
Nitrogen 

Nitrogen 


? 


Blue. 


Indigo. 




Violet. 


Another, ..... 

but closely followed by violet 

hydrogen, .... 

End of band on which violet 

hydrogen is projected, 


1 


58 354 

58 518 

59 208 




Nitrogen 
Hydrogen 






Line beginning another band, 

Faint hazy liue, .... 

Line beginning band, 

Line beginning band, 

Here haze band, .... 

Weak hazy line, beginning a faint 

indistinct band, 
Hazy beginning of a band, 


2-5 
1 

1-2 
1-5 

l» 

0-7 


|!!:. 


59 500 

60 294 

60 480 

61 290 

62 140 

62 690 

63 750 


Marsh-gas ? 


Nitrogen ? 

Nitrogen 
Nitrogen 
Nitrogen 

■i 

? 




Lavender. 



Spectrum ends here, excepting only some ultra-faint ghosts of haze-bands, derived apparently from false reflections. 

Temperature of chamber and apparatus = 64° '0 Far. Spark = 07" long. 

Impurities here are very few and weak. 

The constituents Hydrogen and Nitrogen are abundant ; but the latter seems to tend more towards lines than bands, 
through all the middle of the Spectrum, than in a pure Nitrogen tube. 

This Ammonia tube still further contrasts with the late Alcohol tube, inasmuch as there is here no trace of the combustion 
spectrum of carbon, carbo-hydrogen, &c. ; and we may note at the same time how superior in brilliancy is the Glaucous, to 
the Red, hydrogen line ; a symptom of higher electric temperature here, than in the Alcohol spectrum, wherein the Red was 
the brightest of all the hydrogen lines. 



GASEOUS SPECTRA IN VACUUM TUBES. 

CARBONIC ACID. End-on Tube. October 6, 1879. 

= C0 o . 



113 



Colour. 



Subject of Observation. 



Red. 



Scarlet-Red. 



Light-Red. 



Orange. 



Yellow. 



Citron. 



Green. 



Part 1. 

Faintest band, . . . . 
Faint haze intervenes, . 
Hazy line on red haze, . 
Faint broad haze band, . 
Red line, or sharp beginning of a 
red band 

Part 2. 

Red line, or sharp beginning noticed 

before, .... 
Red Hydrogen, . 

Bundle of thin lines with haze, 



Single line rather hazy, . 

Red band, sharp beginning of, 

fading end of, 
Line terminating much haze, 



Admirably clear, distinct lines, 



Thin line in black space, 

Orange band, intense beginning of, 

weak end of, 
A band of haze just resolvable into 

ultra-thin, close lines, 
Narrow band almost resolvable, 

Line beginning a band of nearly 

resolvable lines, 
fainter end of said band, 

Line, 

Faint line, .... 
Fainter line, .... 
Line, ..... 
Strong beginning of a band of faint 

thin lines, 

Part 3. 

The strong line alluded to before, . 
Band of just resolvable lines extends 

thus far, .... 
A separate thin portion, 

Another separate broader streak, 

Band \ ^ eak fc«™g of, . 
Sharp end of, 



Very faint lines, . 

Citron band \ f lar ? tegLnnmg 
( Faint ending, . 

"Weak lines 

Faint band, begins, 

ends, . 
"Weak band, begins sharply, . 

ends faintly, 



Intensity. 



0-5 
0-3 
1 
0-5 

1-5 



1-5 

5 

1-5 

1 

2-0 
0-5 
1-0 

3-0 
0-5 
10 

2 
1 



2 

1 
1 

1 

1 
2 

0-5 

0-5 

0-5 

4 

1 

1 

0-5 

1 

1 

1-5 

0-5 



Appear- 
ance. 



W.N. Place. 



III::. 



35 878 

37 088 

37 898 

38 352 



38 379 

38 698 

39 308 
39 420 
39 651 

39 858 

40 091 

40 282 
40 622 
40 803 

40 948 

41 086 
41 238 
41 387 
41 498 

41 668 

41 771 

42 049 
42 098 
42 387 
42 525 



42 803 

43 034 
43 138 
43 261 
43 396 
43 526 

43 675" 



43 678. 

44 041 

44 125 
44 272' 
44 363 
44 495' 
44 839, 

44 954 

45 092 
45 185 
45 262 
45 757 

45 884 

46 075 
46 418 

46 866 

47 061 
47 311 



Impurities. 



Hydrogen 

Hydrogen 
Hydrogen 
Hydrogen 
Hydrogen 



Hydrogen 

Hydrogen 
Hydrogen 

Hydrogen 
Hydrogen 



Nitrogen 
Nitrogen 

> Nitrogen 



Hydrogen 



Hydrogen 
Hydrogen + 

43 676 



Hydrogen 
Hydrogen 

Hydrogen 



Hydrogen 
Oxygen 



Constituents 
dissociated. 



Carbon 

Carbon 
Carbon 

r Carbon 

-{ 38 366 

I 

I Carbon 

Oxygen 



Carbon 



Oxygen 



Carbon 



+ Carbon 



Carbon 



Oxygen ? 
Carbon 



Left unclaimed 

for Carbonic 

Acid. 



VOL. XXX. PART I. 



114 



PROFESSOR PIAZZI SMYTH ON 

CARBONIC ACID. End- on Tube— continued. 



Colour. 



Subject of Observation. 



Intensity. 



Appear- 
ance. 



W.N. Place. 



Impurities. 



Constituents 
dissociated. 



Left unclaimed 

for Carbonic 

Acid. 



Green. 



Glaucous. 



Blue. 



Indigo. 



Violet. 



Lavender. 



Part 3 — contimied. 

Hazy lines 

Faint band, 

Fainter haze Hue 

Very dark space intervenes. 
Exceedingly thin line in that space, 

Green band begins sharply ; with 
a weaker spark it is followed by 
a faint Carbo-hydrogen Green 
Giant. Ends in faint haze, 

Faint band, 

Faint hazy lines 



Faintest hazy bands, 

Glaucous Hydrogen, 

Part 4. 

Glaucous Hydrogen, 
Blue band, sharp beginning, 
weak end, . 



Very faint bands, 



Violet band \ Shar P ^g" 111 "^ 
violet Dana, j Faint ending) 

Faintest bands, 

Second ) Sharp beginning, 
violet band, \ Faint ending, . 
Violet hydrogen, . 

Part 5. 

Violet Hydrogen, . 

Faint band, .... 
Faint band, .... 
Faint band 



Stronger grey band, 



1 

1 

1 

0-5 

1-5 

0-3 

0-2 



5 

0-5 

0-3 
1-0 
0-5 
0-5 

1-0 
0-8 
0'5 
0-3 
3 



5 
4 
0-3 

f 0-5 

I 1 

07 
I 0-8 
I 0-5 

3 

0-3 
J 02 
{ 0-2 

2 

1 

2-5 



3 

1 

0-2 

1 

0-2 

1 

0-2 

2-0 
0-3 



j, j 



47 446 
47 646 

47 875 

48 077 
48 338 
48 559 

48 680 



48 841 

49 363 

49 836 

50 139 
50 251 
50 475 

50 774 

51 110 
51 542 

51 784 

52 237 



52 256 

52 510 

53 030 

53 288 

54 207 

54 520 

55 149 

55 620 

56 224 

56 641 

57 069 
57 607 

57 828 

58 256 
58 521 



58 591 

58 996 

59 262 
59 576 

59 961 

60 388 

61 132 

61 670 

62 126 



Hydrogen 



Hydrogen ? 



Hydrogen + 

Hydrogen 

Hydrogen + ? 

) Hydrogen 
[ 52 246 
Hydrogen 



Hydrogen 
Hydrogen 



Oxygen ? 



Carbon 



Carbon 
Carbon ? 



Carbon 
Carbon ? 

Carbon 



Carbon 
Oxygen ? 



Carbon 

Carbon 
Carbon 



End of Spectrum. 

Mems. — Hydrogen and some probabilities of Nitrogen arc the only impurities here. 
This spectrum is very rich in Carbon bands ; but weak in Oxygen, and still more in Carbo-Hydrogen. 
Carbonic Oxide looks the same identical thing as the above Carbonic Acid ; or as explained by Tiialen, the Carbonic Acid 
spectrum is that of Carbonic Oxide, one portion of its oxygen being inert. 

Both Marsh-gas and Olefiant-gas resemble the above in the Carbon bands, but differ by having the lines and the bands of 
Carbo-hydrogen also and strongly. 

The commencing red lines were re-observed on April 7 : 1st at 35 839 ; 2nd at j ^ 2g0 j ; and 3rd at j 3g 122 



GASEOUS SPECTRA IN VACUUM TUBES. 



115 



CAEBONIC OXIDE. End-on Tube. October 13, 1879. 

=CO. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed 
for CO. 


Red. 


Part 1. 


' 0-5 


* | 


35 738 

36 353 


j 


Carbon 






Faint hazy bands, . . . . 


[_ 0-8 


4 


36 613 

37 419 

37 579 

38 170 


1 


Carbon 
Carbon 






Scarlet-Red. 


Part 2. 

Red line, 

Red Hydrogen, .... 

Broad band of barely resolvable 

lines, ..... 

Hazy line, 


1-0 
6-0 

J,, 

07 




38 433 

38 715 

39 103 

39 935 

40 141 


Hydrogen 
Hydrogen 


Carbon ? 
Oxygen ? + 




Light-Red. 






Narrow but solid band, . 


2-0 


I 


40 300 
40 533 


| 


Carbon 






Dark space with faint, close, bright 
lines, 

Very black but narrow space inter- 
venes. 


J 0-7 


mini | 


40 533 

40 878 


> Nitrogen ? 








Pale haze band of resolved lines, . 


1-0 


mill | 
1 

nun | 


40 893 

41 196 








Orange. 


Strong line 

Pale band of resolved lines, . 


2 
1-0 


41 246 
41 321 
41 588 




Oxygen 






Solitary haze line, .... 
Strong orange band — 

Sharp beginning, . 

Faint ending, 
Band begins solid, 
and ends faint in resolvable lines, 

Do. do. 

Do. do. 
Yellow line (not the Salt-line), 

Part 3. 

The yellow line left off with in 

Part 2, 

Faint hazy space with resolvable 

lines, 

Sharp line, ..... 
Strong line, 

Broad hazy space of resolved lines, 
Very narrow band, 


1-0 

3 

1 

2-5 

1 
i 1-5 
| 0-5 
i 1-2 
j 03 

2-0 

|'w 

| 0-5 

10 
2-0 

1-0 
1-0 


1 

mi | 

I 

llll 

■ 1 


. 41 655 

41 784 

42 058 
42 128 
42 432 
42 454 
42 665 

42 760 

43 034 
43 142 

43 140 

43 233 
43 465 
43 534 
43 706 

43 768 

44 140 
44 339 
44 376 


1 Nitrogen ? 
> Nitrogen ? 
j Nitrogen ? 

( Hydrogen 
f 43 141 

Hydrogen 
| Hydrogen 


Carbon 


? 


Yellow. 






Broad space of just resolvable haze, 


2-0 


■ | 


44 502 

44 883 


1 




f 


Citron. 


Dark space with faintest close lines, 


0-5 


1 


44 967 

45 192 










Grand Citron band — 
Sharp beginning, . 
Faint end, .... 
Second still fainter end, 


4 
1-0 
4 

) 15 


I*. | 


45 281 

45 855 

46 392 

47 124 


1 

s 


Carbon 










Carbon 




Green. 


Group of faint lines and haze, 

Haze band, 

Faint lines in haze, 


f 0-7 
( 1-0 
) 1-0 

1-5 

| 1-0 
i 1-0 


i 

1 , 

.4 j 

1 

1 


47 435 
47 677 

47 884 

48 202 
48 385 
48 503 
48 721 


Hydrogen 

\ 


Oxygen ? 


? 



116 



PROFESSOR PIAZZI SMYTH ON 

CARBONIC OXIDE. End-on Tube— continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed 
for CO. 


Green. 


Part 3 — continued. 

Grand beginning of green band, 


5 




48 842 




Carbon 






Faint show of Green Giant, . 
Tail of Green band extends thus far, 
Haze band, 

Hazy lines 


2 

0-5 

1-5 

( 1-2 
\ 07 
( 1-2 


1 

: i 


49 201 | 

49 877 ) 

49 953 } 

50 140 ) 
50 235 
50 380 
50 492 


Carbo- 
Hydrogen 

Hydrogen ? 


Carbon 






Glaucous. 


Resolvable haze band, . 

Faint resolvable I faint beginning, 

haze band, ( sharp end, 
Hazy line, ..... 
Very faint hazy space follows, 
Glaucous Hydrogen, 


I 0-5 
{ 1-0 
j 0-5 
| l'O 

1-0 
0'4 
5 


i : 

! 1 

B 


50 567 ) 
50 862 \ 

50 964 

51 263 

51 505 

52 256 ( 


Hydrogen x 

i 
Hydrogen 


Carbon ? 
Carbon ? 


? 




Part 4. 

Glaucous Hydrogen, 
Grand blueband ; sharp beginning of, 
Do. do. weak end of, 


4 
3 
0-5 

\\ 

\ 0-5 
0-5 

I 1 
2 
1 

[ 0-3 

\ 1-0 

J ,5 

0-3 

2-0 


i0 


52 260 ( 

52 523 ) 

53 008 \ 
53 775 


52 258 
Hydrogen 


Carbon 






Faint hazy lines, in still fainter 
haze, 

Violet band ; sharp beginning, 
Do. faint ending, 

Hazy lines, 

Second violet band; sharp be- 
ginning, 

Do. do. faint ending, . 
Violet Hydrogen, .... 


! 1 1 
| ■ | 

1 


54 220 

54 749 

55 177 

55 514 

56 240 ) 

56 652 \ 

57 037 
57 512 

57 807 ) 

58 235 ) 
58 512 j 


Hydrogen ? 
Hydrogen ? 

Hydrogen 


Carbon 
Carbon 




Blue. 


Indigo. 




Violet. 


Part 5. 

Violet Hydrogen, .... 
Faint band ; sharp beginning of, . 
Do. faint end, . 

Do. do. . 

Hazy line, ..... 

Faint band, 


1-5 

1-0 

3 

| 10 

| 0-3 

0-5 

| 1-0 

( 0-3 


Itl 


58 510) 

58 817 

59 083 \ 
59 376 

59 760 

60 211 

60 410 I 

61 028 f 


58 511 
Hydrogen 


Carbon 
Carbon 






Lavender. 


Rather stronger and greyer band, . 

Very faint band 

Suspected line, .... 


1-5 

j 0-4 

0-5 

o-i 


U i 


61 582 | 

62 017 i 

62 716 

63 689 


? 


Carbon 
Carbon ? 





End of visible Spectrum. 

When viewed with higher dispersive powers as 11°, 22°, and 33° Dispersion A to H, the large bands of this speclrum, 
however solid, smooth, and compact they may appear to small dispersion, separate in the most marvellous manner into thin 
hard separate and as well-defined lines as any we could wish to see ; but it would take weeks to measure them all. 

Meanwhile, with this spectrum, Hydrogen is the only large and certain impurity ; Carbon is the chief, and Oxygen a minor, 
constituent. The yellow unnamed lines 43 141 and 43 706 appear ; also the doubtful 51 505, to be low temperature 
Hydrogen lines. 



GASEOUS SPECTBA IN VACUUM TUBES. 



117 



CHLORINE. January 16, 1880. 



CI. 



Colour. 



Crimson- 
Red. 



Red. 



Scarlet-Keel. 



Light-Ked. 



Orange. 



Yellow. 



Citron. 



Green. 



Subject of Observation. 



Part 1. 

A certain, but thin and sharp 

line 

Band, 

Strong red line, probably 

Red Hydrogen, . 

Part 2. 

Strong Eed Hydrogen, . 

Faint group of lines or haze, . 
Hazy band of lines, 

Thin line 

Faint lines in a still fainter 

illuminated region, 
Very thin line, terminating 

faint haze, .... 
Faint line or group of lines, 

Distinct line, black space on 
either side, .... 
Do. do. 

Do. do. 

Faint broad flat haze band, 
Band of almost resolved lines, 
Do. do. 

Part 3. 

Thin line in faintest haze, 
Line in hazy region, 

Lines bounding a faint band, 

Very faint hazy line, 
Line, .... 
Very faint line, 
Faint band, . 



Exceedingly brilliant line in 

black space, 
Another do. do. 

Another do. do. 

Another do. do. 

Faint line, .... 

Faint line 

Faint line, .... 

Brilliant solitary line, shown 
by higher dispersion to be 
double ; its true place seems 
on re-examination to be 
rather smaller, say 48 685, 
and by no means increased 
like the tabular 6" spectral 
place, which I suspect there- 
fore to be in error, 



Intensity. 



1 
1-5 

5 



0-4 
1-0 

0-5 

0-3 
1-0 
0-3 
0-5 



5-5 
6-0 
4-0 

1 
1 

1 



Appear- 
ance. 



W.N. Place. 



34 870 

37 963 

38 702 

38 671 

38 956 

39 398 

39 677 

40 528 

41 046 
41 210 

41 375 

41 557 
41 714 

41 816 

42 536 
42 681 

42 867 

43 052 
43 279 



43 475 

43 862 

44 205 

44 379 

45 078 
45 325 

45 560 

46 005 



46 548 

46 686 

46 869 

47 146 

47 614 

47 902 

48 109 



48 716 



Impurities. 



Chlorine Spectrum 
from 6-inch spark 
by other observers. 



1 Hydrogen 
I 38 687 
J Hydrogen 



Nitrogen ? 
Hydrogen ? 

Hydrogen ? 



Oxygen ? 
? 

2 

Hydrogen ? 

Nitrogen ? 
Nitrogen ? 

Nitrogen ? 



Inten- 
sity. 



2 
2 

2 
2 
2 
2 
2 
2 
2 

10 

10 
10 

10 

2 
2 
4 



W.N. Place. 



37 472 

37 955 

38 110 
38 223 



41 584 



42 675 
42 804 



43 882 

44 436 

44 679 

44 765 

45 035 
45 347 
45 543 
45 848 

45 905 

46 520 

46 657 

46 847 

47 167 

47 512 

47 699 

48 160 



48 733 
48 799 



Left unclaimed 
and probable 
for Chlorine 
in a 1-inch 
spark only. 



34 870 



41 375? 

41 557? 
41 714? 



43 862? 



45 078 
45 325? 

45 560? 



46 548 

46 686 

46 869 

47 146 

47 614 ?j 

47 902? 

48 109? 



48 716 



118 



PROFESSOR PIAZZT SMYTH ON 

CHLORIN E— continued. 















Chlorine Spectrum 
















from 6-inch spark 


Left unclaimed 








Appear- 
ance. 






by other observers. 


and probable 


Colour. 


Subject of Observation. 


Intensity. 


W.N. Place. 


Impurities. 




for Chlorine 
in a 1-inch 


















Inten- 
sity. 


W.N. Place. 


spark only. 




Part 3 — continued. 


















Faint line, .... 


1 


1 


48 964 




2 


49 033 






Faint line 


1 




49 128 




2 


49 072 






Do 

Stronger line, 


1 
2 


1 , 


49 295 
49 824 




2 

2 


49 215 
49 225 


49 824 






Do. .... 


2 


| 


50 054 




2 


49 319 


50 054? 




Line, ..... 


1 


1 


50 872 




2 


49 338 


50 872 ? 


Glaucous. 




• 








4 
6 
6 
1 
1 
2 
2 
4 
4 
2 
2 
4 


49 793 

49 813 

50 030 
50 138 
50 358 

60 738 
50 759 

50 818 

51 065 
51 335 
51 396 

61 520 






Line with haze before it, 


1-5 


1 


51 644 




4 


51 584 


51 644? 




Line, ..... 


2 


| 


51 776 




6 


51 762 


51 776? 




Line 


2 


1 


51 860 




6 


51 845 


51 860? 




Merest suspicion of Glaucous H. 


o-i 


1 


52 244 


Hydrogen 








Glaucous. 


Part 4. 




















I 4 


I 


52 671 




10 


52 642 


52 671 




Triplet of most notable lines, . 


4 


1 


52 804 




10 


52 762 


52 804 






i 4 


I 


52 960 




10 


52 916 


52 960 






f 0-5 
1 0-5 


1 


53 160 




2 


53 025 








1 


53 606 




6 


53 070 






All faintest lines with suspi- 


J 1-0 


1 


53 901 




1 


53 115 






cions of bands to some, 


1 0-5 




54 660 




2 


53 160 






0-5 


I" - 
1 


55 123 




6 


53 170 




Blue. 




I ro 

( 0-5 
1'0 


1 

1 


55 548 

55 927 

56 477 




2 
8 
1 


53 305 
53 484 
53 917 


55 548? 


Indigo. 


Do. do. 


i-S 


1— 


57 414 




2 


54 624 








I 1-5 


.1. 


58 327 




2 


54 813 














2 


55 039 
















2 


55 337 
















-1 


55 469 
















55 530 






Tube tested and found quite 
cool, 


i 








|» 


58 445 




Violet. 


Part 5. 


















Group of lines in haze, . 


2 


lilll 


58 456 




2 


58 551 


58 456? 




Hazy line 


0-3 




58 935 




4 


58 934 


58 935? 




Hazy band, .... 
Hazy band, .... 


2 

2 




59 554 




2 


59 165 
59 343 


59 554? 






60 646 




1 




Fainter band, .... 


1 




61 459 




1 


59 386 




Lavender. 


Very faint band, 
Do. do. 
Do. do. 


0-3 
0-3 
0'2 




62 236 

62 686 

63 574 




4 


59 651 








End of Spectrum. 

None of the powerful bands of Carbon are here, as they are eminently in Hydrochloric Acid. 

The Hydrogen lines here are weak as impurities only, compared with what they are in Hydrochloric Acid, where they are 
a constituent. But the Chlorine lines are far stronger here. 

Chlorine, however, is peculiarly absent from all the other tubes, absolutely too if tested by its chief line, the close double 
48 716. 

The column of 6-inch spark, or high -temperature, Chlorine lines is derived from Thalens, Huggins, and others who have 
so observed, and is curious for both its agreements and disagreements with my low-temperature, small spark, spectrum lines. 



GASEOUS SPECTRA IN VACUUM TUBES. 



119 



CYANOGEN. Old Tube. 

Symbol =Cy or CN. Carbon and Nitrogen. 

An old and rather used-up example ; so that its light is no longer white but pink ; and the inside of the 

bulbs hazy. Observed July 8, 1879. 



Colour. 



Eed. 



Scarlet- 
Red. 



Light- 
Red. 



Orange. 



Yellow. 



Citron. 



Subject of Observation. 



Visions of bands, like those which fol 
low, but only of . . . 

Line, faint, .... 

Band, first and faintest of a very regu 
lar series, .... 

Another and brighter, 
Do. do. 

Do. do. 

Red Hydrogen (the merest trace), 
Band again, . . 

Do. .... 

Do. .... 

Do. .... 

Do. .... 

Do. .... 

Do. .... 



Intensity. 



o-i 

1 

2 

2-5 
3 
3-5 

0-3 
3-5 

3-5 
3-2 
2 7 
2-5 



1-3 



Appear- 
ance. 



W.N. Place. 



36 250 

36 720 

37 080 
37 228 
37 620 

37 823 

38 165 
38 302 
38 612 

38 712 

38 765 

39 015 
39 230 
39 500 

39 624 

39 980 

40 115 
40 445 

40 580 

40 935 

41 038 
41 365 

41 465 
41 705 



Impurities. 



Hydrogen 



(No 
Oxygen) 



Constituents 
dissociated. 



Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 

Nitrogen 

Nitrogen 

Nitrogen 

Nitrogen ? 

Nitrogen ? 



Unclaimed 
features left for 
Cyanogen, Old. 



This last band faintest of the set of 11 similar breadth bands, also split down the middle by a black line ; 
afterwards begins a new and bright series of 7 similar breadth bands ; as in Thalen's Nitrous Oxide, by 
others called Nitrogen. 



Very strong, well-defined band, 



Do. 
Do. 
Do. 
Do. 
Do. 



do. 

do. 
do. 
do. 
do. 



Last band of this set, 





4 


. | 


41 805 

42 145 




4 


. | 


42 230 
42 570 




4 


ii| 


42 610 
42 902 




4 


. | 


42 974 

43 242 




4 


. j 


43 310 
43 640 




4 


.] 


43 415 

44 010 




4 


. | 


44 090 
44 380 



Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 



These two series of most remarkable and regular bands in Cyanogen are still more like Thalen's Bi-oxide 
of Azote, other's Nitrogen, than the Nitrogen tube itself. The Hydrogen has only lately come into view. 



120 



PROFESSOR PIAZZI SMYTH ON 

CYANOGEN. Old Tv be— continued, 



Colour. 



Subject of Observation. 



Citron. 



Green. 



Glaucous. 



Blue. 



Indigo. 



Violet. 



Lavender. - 



A black space follows, taint line there- 
in, developes afterwards into a cyan- 
ogen line 

Sharp edge of a citron band, 

A band after a region of thin close lines, 
Very sharp line begins a band of lines, 

Green band begins, .... 
A line in that band, .... 

An isolated line, .... 



Intensity. 



Appear- 
ance. 



W.N. Place. 



44 878 

45 200 

47 806 

48 582 

48 862 

49 050 

49 350 



Impurities. 



Constituents 
dissociated. 



Carbon 



Nitrogen ? 



Carbon 

Nitrogen ? 

Nitrogen ? 



Unclaimed 
features left for 
Cyanogen, Old. 



Cyanogen 
44 878 



\ Cyanogen 
48 582 



J Cyanogen 
| 49 350 



N.B. — The above two lines are certainly not Blow-pipe's, or Carbo-hydrogen's green-giant and its second 
following line, which read as now measured specially in a flame close by, 49 178 and 49 516. 



Faint band begins, 

Stronger band with central line, 

Line, 

Line, 



Line 

Fainter line, .... 
Sharp beginning of graduated band, 

Glaucous Hydrogen, . 

Very faint haze band, 

Sharp beginning of strong band, 

Faint line, ..... 

Sharp beginning of strong band, 

Faint line, 

Band begins, .... 

Line, ...... 

Sharp beginning of strong band, 
Very weak band, 
Graduated band begins, 
Do. do. 

Faint band begins, 

Strong linear beginning of a band, 

Grand line peculiar to Cyanogen, 

Strong beginning of a band, 

Very faint band, 

Band begins 

Faint band begins, 

Very faint do. ... 

Stronger band begins, 

Faint band begins, 



2 
3 
3 
3 

3 
2 
4 

2 
2 
4 
1 

4 

1 

3 

2 

4 

0-5 

3 

3 

1 
3 

5 

4 

2 

3 

2 

1 
3 
2 



49 996 

50 170 
50 510 

50 728 

51 100 
51 260 

51 650 

52 250 
52 585 

52 794 

53 836 

53 960 

54 460 

54 642 

55 271 

55 630 

56 370 

56 645 

57 602 

58 290 

58 530 

59 405 

59 511 

60 015 

60 535 

61 335 

62 065 

62 625 

63 496 



Hydrogen 



Nitrogen ? 
Nitrogen ? 
Nitrogen ? 

Nitrogen ? 
Nitrogen 



Carbon 
Nitrogen 



Nitrogen 

Nitrogen 
? 

Nitrogen 

Carbon 

Nitrogen 

Nitrogen 



Nitrogen 

Nitrogen 
? 

Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 
Nitrogen 



Cyanogen ? 



Cyanogen 
53 963 



I Cyanogen 
\ 59 405 

( Cyanogen 
j 60 015 



End of Spectrum. 

This Cyanogen tube, remarkable at first, for its bright white light to the eye, — so that its end-on view reminded me of a 
little full Moon high in the sky, — possessed then large carbon bands, no hydrogen at all, most brilliant lines and bands in the 
violet, and a more regular vertical stratification of what are here called Nitrogen bands, in the red, orange, and yellow, than 
what the Nitrogen tube itself showed. 

Now the light is pink, hazy, and faint. The Carbon bands far smaller ; a little hydrogen has come into view ; the extra 
regularity of the Nitrogen bands being still partly kept up. 

The powerful violet line at 59 405, may become useful as a reference for place to many observers. 



GASEOUS SPECTRA IN VACUUM TUBES. 121 

CYANOGEN. Very Old Tube (End-on). February 28, .1880. 

(CN. Carbon and Nitrogen.) 

Six mouths further change are seen here on the " Old Cyanogen tube" observed iu July 1879 ; Hydrogen 
impurity has grown largely. Carbon bands have notably decreased, also those of Nitrogen ; but special Cyanogen 
lines are stronger and more numerous. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W. N. Place. 

i 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 

for Cyanogen 

(very Old). 


Red. 


Part 1. 

Three faint Nitrogen-like bands, . 

Part 2. 

Nitrogen-like band, 

Red Hydrogen (grown ! !) 

Nitrogen-like bands, 


( 0-2 
\ 0-4 
( 0-5 

1-5 
3-5 

r i-5 

1-6 
2-0 
2-5 
2-5 


... 

m j 


36 877 

37 596 

37 991 

38 455 
38 727 

38 934 

39 349 
39 660 

39 970 

40 123 
40 451 
40 580 
40 891 


Hydrogen 

1 
i 


Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 

Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 




Scarlet-Red. 


Light-Red. 






Strong line followed by a band, 


2D 


Hi::. 


41 146 


j Oxygen ? 
\ Hydrogen? 

Hydrogen ? + 

1 




2 


Orange. 


Another line, haze, 
and then a very black space. 

Very firm band, .... 
Do. do 


2-0 

3 

2-5 


|!:=. 


41 552 

41 810 ) 

42 156 \ 

42 251 
42 575 


( Carbon? 
< and 

( Nitrogen ? 

Nitrogen 


Cyanogen 




Do., but fainter, 


1-5 


■ ! 


42 622 
42 904 


[ 


Nitrogen 






Salt line apparently ! 


3 


1 


43 109 










Yellow. 


Part 3. 

Salt line repeated, 


3 

r i-o 


1 


43 109 
43 285 
43 619 


> Sodium 

1 


Nitrogen 






Three Nitrogen-like bands, . 


\ 1-5 


m \ 


43 735 

44 017 


i 


Nitrogen 






Citron. 


Thin line in a black space, 

Very remarkable group of clean 
graduated lines in black space, . 


L 2-0 
0'4 

( 2-0 

1-0 

( 0-3 


m | 


44 099 

44 348 
44 491 

44 706 
44 841 
44 946 


i 

Hydrogen ? 


Nitrogen ? 


( True 
1 Cyanogen 
{ group 
j beginning 
L 44 706 




Remains of Carbon's Citron band, . 


) 2'0 
\ - 3 


!, j 


45 293 
45 469 


1 
1 

t 

1 ! 


Carbon 






A band in a now beginning hazy 

region, 

Do. do. 

Band of haze, almost resolvable 
into lines, 

Line in hazy region, 

Broad band of just resolved lines, . 


J,. 

1-0 

},0 

1-0 

1-0 


".j 

:i: 

liimil j 


45 672 

45 940 

46 322 
46 608 

46 879 

47 314 
47 792 


Nitrogen 
Nitrogen? 
Nitrogen 

? 

? 








Very black space follows. 




t 










Grand clean green line, . 


3 


1 


48 562 






( Cyanogen 
\ 48 562 


Green. 


Remains of Carbon's Green band, . 

Black space in midst of which Green 
Giant should be, but it is abso- 
lutely not. 


2-0 


M 


48 846 

49 135 


I 


Carbon 





VOL. XXX. PART I. 



122 



PROFESSOR PIAZZI SMYTH ON 

CYANOGEN. Very Old Tube (End-on) — continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W. N. Place. 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 

for Cyanogen 

(very Old). 




Part S — continued. 














Green. 


Grand line followed by faint haze, 
Very black space follows. 


2-5 


L' | 


49 335 
49 526 






) Cyanogen 
\ 49 335 






System of clean lines, but utterly 


( 2 


) ( 


50 019 ) 






( Cyanogen- 




altered if condensed spark be 


{ 1 


I "'1 


50 191 } 


Hydrogen ? 


! 




employed, ..... 


( 0'3 


50 287 ) 






1 group 




Black space follows. 
















Inverted band ending sharply, but 


|. 














changed by condensed spark into 


\4 | 


50 629 | 
50 779 \ 


Hydrogen + 


1 


( Cyanogen- 




a double line, the brightest of the 


2 




( band 




whole spectrum, 


) 














Black space follows. 
















Group of very thin lines, 


1 


„,j 


51 117 
51 275 


? 








Very thin line. .... 


0-5 


1 


51 465 


Hydrogen ? 








Hazy and thin line, 


0-6 




51 707 


i 


Nitrogen ? 




Glaucous. 


Signal line, clear and sharp ; not 

altered by condensed spark, 
Another similar, supposed Glaucous 


!• 


1 


52 019 


? 








Hydrogen, not altered by con- 


| 3 


1 


52 257 


Hydrogen 








densed spark, .... 


J 














Part 4. 


f 0-2 
1-0 


1 


52 282 
52 423 






Cyanogen 




Perfectly sharp lines in black space, 


2-0 

0-8 

0-3 

L 0-3 


1 

1 
1 


52 863 

53 328 
53 659 
53 795 




Nitrogen ? 


( Cyanogen 
\ 52 863 




Signally bright line followed by haze 


3 


h \ 


53 963 | 






( Cyanogen 




(not Nitrogen nor Carbon), 


0-2 


\ L \ 


54 260 | 






\ 53 963 






1 07 


l 


54 570 






Cyanogen ? 






i 


54 694 










Clear lines in black space, 


J 07 
1 1-0 


i 


54 796 

55 050 


Hydrogen ? 








Blue. 


Haze band, ..... 

Do 

Do 


1 0-8 
I 0-5 

0-3 
l'O 
1-3 


i 


55 189 
55 310 

55 782 

56 589 

57 354 


Hydrogen ? 
















Indigo. 


Thin line, ..... 

Haze band, ..... 

Do 


1*0 

0-7 
0-5 


::: i 


57 744 

57 987 

58 159 












Oxygen ? 








Hazy line (perhaps faint violet H), 


l'O 


.:!:. 


58 461 


Hydrogen 








Hazy band, ..... 


0-5 




58 613 










Very black space follows. 














Violet. 


















Part 5. Black space. 














Grandly strong violet line, followed 
by a band, specially characteristic 
of Cyanogen, .... 


/ 3-5 

i 0-3 


jl, J 


59 389 ) 
59 646 \ 






( Cyanogen 
( 59 389 




Line begins a band, 


\ 2 
\ 0-2 


1 *•! 


59 961 ) 

60 356 \ 


■i 




Cyanogen? 






Line 


l'O 


i 


60 541 




Nitrogen ? 






General faint haze follows, 
















Hazy line, 


l'O 




61 284 




Nitrogen ? 




Lavender. 


Black space follows. 
















Line begins a haze band, 


1-5 


1::. 


62 691 




Nitrogen ? 






Faint haze band, .... 


0-3 




63 808 










Very faint haze, .... 


o-i 




64 836 









End of Spectrum. 



GASEOUS SPECTRA IN VACUUM TUBES. 



123 



CYANOGEN, Another example of, long laid by unused. 

October 10, 1879. 

= CN. 



End on Tube. 



This tube gives an eminently brilliant and white light to the eye ; just as did the former old and much- 
used tube, before it went wrong, became pink in colour, pale and hazy, besides apparently sparking from the 
polar wires inside their glass fixings. In the spectroscope this tube showed 12 grand bands of Carbon, and 41 
lesser of Nitrogen, its proper constituents ; also some Hydrogen, and also Carbo-hydrogen, or Blow-pipe flame, 
lines as impurities, but no Oxygen ; and has not been deemed worthy of being printed at length. 



HYDROCHLORIC ACID. January 23, 1880. 
HCl=Hydrogen Chloride. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities 


Constituents 
dissociated. 


Left unclaimed 
for Hydro- 
chloric Acid. 


Crimson- 
Red. 


Part 1. 

Very clear and distinct line, . 
Broad faint haze, .... 

Strong band, .... 
Fainter band, .... 


1-5 

0-5 

1-5 
1-0 


1 

m | 
■ 


34 960 

36 162 

37 049 
37 429 

37 628 

38 123 


Carbon 
[ Carbon 

[ Carbon 


Chlorine 




Red. 




Scarlet-Red. 


Line or band just before Red Hydro- 


j,. 




38 414 


"1 Carbon 
1- 38 419 
J Carbon 

Oxygen? and 
Nitrogen 








Part 2. 

The line or band beginning just 

before Red Hydrogen, 
Red Hydrogen, .... 

Bundle of lines in haze, 

Thick broad or multiple line, 


!• 

5 

w 

2 


■ 

! - 

III 


38 424 

38 742 

39 352 I 
39 506 ( 
39 715 


Hydrogen 




Light-Red. 




Band, 

Band, 

Band, 


1-5 
1-5 
1-5 


,,\ 

■■■■■■ < 

.j 


40 035 
40 246 
40 377 
40 607 
40 676 
40 899 


| Nitrogen ? 
| Carbon ? 
1 Nitrogen ? 








Orange. 


Broad line, ..... 

Strong line, 

Strong line, 

Faint narrow band, 
Faint bundle of lines, 
Strongest line in a faint bundle 
accompanied by a Red band, 

Thin line, 

Strong line among thin ones, 
Close thin lines intervene, 
Strong central bundle, . 
Close thin lines intervene, 


2-0 
2-5 
2-5 
1-3 
1 

| 2-5 

0-7 
2-5 

2 


i 

i 

iiiii 
urn 

inn 


41 022 
41 291 
41 466 
41 588 
41 731 

41 920 

42 029 
42 187 

42 540 


? 
Oxygen ? 

Carbon ? 
Carbon'? 


Chlorine 

Chlorine 

Chlorine 

+ Chlorine 

Hydrogen ? 
Chlorine ? 


2 




Bundle of close lines, 


2 


mi! 


42 784 




( Hydrogen? 






Thin line, 

A yellow line, slightly beyond 
Sodium, &c. , . 


1 


i 
i 


42 919 

43 116 


1 


Hydrogen ? 


2 




Yellow. 


Part 3. 

Bundle of thin lines, after last ter- 
minal line, .... 

Very thin line, .... 

Stronger line, .... 

Exceedingly sharp, bright intrinsi- 
cally line, 

Intervenes a space, full of sharp 
lines, 


| 07 

0-3 
1-0 

!'• 

j 0-3 


llll 

1 

1 

1 

llll 1 


43 303 

43 445 
43 506 

43 702 

43 721 

44 114 


\ 


Chlorine ? 

Hydrogen ? 
Hydrogen 

Chlorine? 


2 





124 



PROFESSOR P1AZZI SMYTH ON 

HYDROCHLORIC ACID— continued, 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 
for Hydro- 
chloric Acid. 




Part 3 — continued. 

Terminal line of that space, . 

Narrow band of lines, . 


1-5 
2 


1 


44 133 
44 277 
44 395 


i 

1 


Hydrogen 
Hydrogen + 






Dual lines in a region of linelets, . 


( 1-5 
\ 15 


1 "| 


44 549 
44 696 


I 


Hydrogen 




Citron. 


Single line in a region of fainter 
lines, 

Single line in do. do. — perhaps 
Citron 1, of Blow-pipe flame, 


| l'O 
1-5 


:i: 


44 920 

45 113 




Chlorine 






Grand Citron band, sharp beginning 

of 

Line on tail of above, 

Line still on tail of Citron band, . 

Line, after a region of linelets, 

Line, 

Line in hazy region, 

Lines in region of half-resolved 
linelets 

Peculiar nebulous band, amongst 
linelets, 

Strong line, after half-resolved 
linelets, ..... 


1 

2 

1 

1-5 

2-0 

( 1-5 
) 0-5 
) 0-5 
( 03 

| 0-5 
| 2-0 


::.. 


45 250 


Carbon 








1 

1 iili! 

1; 

in 

ill 
:i: 

1 


45 508 

45 930 

46 571 
46 663 

46 854 

47 118 
47 407 
47 666 

47 919 

48 272 
48 662 


? 


Chlorine 
Hydrogen ? 
Chlorine 
Chlorine 
Chlorine 
Chlorine 

Chlorine ? 
Chlorine ? 

Chlorine 




Green. 




Grand green Carbon band, 
Blow-pipe's green giant on tail of 
band, ..... 
Line 

Line, possibly green giant 2, . 

Line, ...... 

Sharp beginning of faint band, 

Do. do. 

Do. do. 

Do. do. 
Possibly a dual line in faint haze, . 


5 

j 2 '5 

0-5 

1-0 

1-5 
0-7 
0-5 
0-5 
1-0 
1-5 


1 

'1 

1 
1 


48 840 

49 175 
49 396 
49 538 

49 789 

50 202 
50 584 

50 791 

51 008 
51 468 


Carbon 
( Carbo- 
( Hydrogen 

( Carbo- 
( Hydrogen 


Hydrogen ? 
Chlorine ? 
Chlorine ? 






Glaucous. 


Brilliant Glaucous Hydrogen, 
Blue band begins, .... 


6-0 
4 


i„ 


52 229 
52 502 


Carbon 


Hydrogen 






Part 4. 

Exceeding sharp, thin line, . 
Sharp line, ..... 
Sharp distinct line, 

Paint band, ..... 

Band begins sharply, 

ends less sharply, 
Line, ...... 

Band 

Sharp line, probably Violet Hydro- 
gen, 


1-5 

2 
2-5 

0-7 

}• 

0-5 

1-0 


i 

-i 

i 
1 


52 721 
52 821 
52 983 

55 488 

55 804 

56 334 

56 816 

57 557 

58 012 

58 538 


f 

[ Carbon 
Carbon 


Chlorine 
Chlorine 
Chlorine 

Chlorine 
Hydrogen 




Blue. 


Indigo. 




Violet. 


Part 5. 

Line, supposed Violet Hydrogen, . 
Line, fainter, .... 

Band, 

Line, ...... 

Do 

Do 

Faint line or band, 

Faint band, ..... 

Last certain light, 


2-5 

1-5 

0-5 

1-0 
1-5 
5 
1-5 
0-3 

o-i 


1 

i 


58 514 

58 936 

59 440 

59 669 

60 289 
60 530 

60 823 

61 675 

62 618 

63 942 


| 58 526 

Marsh-gas 
Marsh-gas ? 
Marsh-gas ? 
Marsh-gas? 


Hydrogen 

Chlorine 

Chlorine 




Lavender. 



January 26. — After much trying of various prisms on this tube, I was struck to-day with its light being now generally 
pink in place of blue, — with the standard Hydrogen lines being very strong, — and with large groups of lines in the Red, 
Orange, and Yellow, being like those in a Hydrogen lube; also like those produced in old Cyanogen tubes, by too abundant 
practice and use; while the Chlorine lines are becoming marvellously fault ! and the Carbon bands are nearly gone! The 
Carbon may very probably be burned and deposited out; but is the Chlorine changing into Hydrogen? 



GASEOUS SPECTRA IN VACUUM TUBES. 

HYDEOGEN. End-on Tube. August 23, 1879. 



125 



Colour. 



Crimson - 
Red. 



Red. 



Scarlet-Red. 



Light-Red. 



Orange. 



Yellow. 



Citron. 



Subject of Observation. 



Part 1. 

Sharp line, . 

Hazy band, . 

Faint haze begins, 
Line in haze, 
Haze band, . 
Haze band, . 



Red Hydrogen, excessively bright 

Part 2. 

Red Hydrogen repeated, 

Broad haze band, . 

Line in the haze, . 

Second line, . 

Haze intervenes, . 

Line, 

Haze intervenes, . 

Line, 

Broad line, . 

Narrow line, . 

Broad and hazy line, 

Broadest line yet, . 

Very black space intervenes. 
A line, .... 
A fainter line, 
Brightest line yet, 

Double line, . 

Very black space intervenes. 

Broad band with bright line in 
middle, .... 

Broadest cum brightest line yet, 

Thin line, .... 

Band of closely packed thin lines, 

Close double line, . 
Single line, .... 
Very dark space intervenes, 
Strong line, .... 

Part 3. 

Yellow line of last part, 
Faint double or treble line, . 
Faint double, 

Thin but sharp and strong line, 
Brightest line yet in this part, 

Band of infinitely fine lines, with 
a stronger near the middle, 

Strong line, .... 
Very black space intervenes, 

Hard band or thick line, 

Very thin line, 
Thin bright line, . 
Haze of fainter lines intervenes, 
Thin bright line, . 
Haze of fainter lines intervenes. 
Thin bright line, . 
End of the fainter intervening line 
haze, 



Intensity. 



0-3 

0-2 

0-1 
0-3 
0-3 
0-3 

10 



10 
1 
2 

1-5 
1 

1-5 
07 
1-5 
3 
1 
2-5 



2-5 
1-5 
4 

1-5 



2 
3 
2 

4-5 

1 

2 

2 

1 



3 
1-5 

1 
2 
4 



3'5 

0-5 
2 



1 
0-5 



Appear- 
ance. 



W.N. Place. 



32 616 

33 020 

34 200 

35 602 

36 230 

37 090 

37 736 

38 707 



38 707 

39 000 
39 314 
39 466 

39 666 

39 885 

40 114 
40 286 
40 489 
40 666 
40 822 

40 968 

41 120 
41 225 
41 354 
41 510 

41 704 

41 812 

42 010 
42 060 
42 210 
42 302 
42 401 
42 609 
42 810 

42 950 

43 148 



43 


151 


43 


310 


43 


430 


43 


545 


43 


698 


43 747 


43 


862 


44 000 


44 112 


44 225 


44 


363 


44 


455 


44 547 


44 


693 


44 


900 


45 


130 



Impurities. 



Oxygen 

Nitrogen 

Carbon 
Oxygen 
Carbon 
Carbon 



38 707 
Oxygen 



6" condensed 
spark Tables. 



38 707 



Oxygen 



Carbon, &c. 



43 150 



44 294 



Left for low 
temp. Hydrogen 



38 707 



39 466 
39 666 

39 885 

40 114 
40 286 
40 489 

40 744 



40 968 

41 120 

41 354 

41 510 



41 812 

42 135 

42 810 

43 150 



43 
43 


544 
698 


44 


112 


44 


294 


44 


547 


44 


693 



Now begins a Citron-green region with background of light resolvable haze more evenly distributed than 
before, so that there is no more of the very black interspaces previously noted. 



126 



PROFESSOR PIAZZI SMYTH ON 

HYDROGEN. End-on Tube— continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


C" condensed 
spark Tables. 


Left for low 
emp. Hydrogen. 


Citron. 


Part 3 — continued. 

A band, 

Haze just resolvable continues up 
to this line 

Resolvable haze of thin lines inter- 
venes. 

A distinctly double line, 

Rather strong lino, 

A haze of thin close lines intervenes, 

Stronger line, .... 

Line haze intervenes, 

Evident line in the haze, 

Haze of lines intervenes, 


2 

2 

1-6 

07 
2 

1-5 


. ! 
"J 

I 
I 


45 268 
45 406 

45 893 

46 748 

46 880 

47 143 

47 668 
47 920 


[ Carbon 

Oxygen 

? 

Oxygen 




45 893 
47 920 




Green. 


Hard haze band, .... 


1-5 


| 


48 190 
48 390 










Haze of lines intervenes, 
Sharp beginning of a grand green 
baud, 


1* 




48 860 


Carbon 








Line just visible in that band's tail, 

Another fainter, .... 
Decreasing haze of lines intervenes. 
Beginning of a band of hazy lines, 
Stronger lines than others, 


2 
1 

2 
2-5 


1 


49 180 
49 380 

49 970 

50 649 


I Carbo- 
( Hydrogen 




49 380 

49 970 

50 649 






Band of faint lines, 


1-5 


-1 


50 982 

51 278 


i ? 






Glaucous. 


A close hazy double line ; subse- 
quently found cleaner and clearer, 
and simply a double line, . 

Glaucous Hydrogen, painfully bright, 


V 

12 


■ 


51 463 

51 537 

52 254 


) 




51 463 
51 537 




Part 4. 








V 52 252 


52 255 


52 252 




Glaucous Hydrogen, 

Sharp beginning of band, graduated 
band, ..... 

Faint line in haze, 

Another rather stronger line, 

Sharp line, strongest between 
Glaucous Hydrogen and Violet 
Hydrogen ; its place elsewhere 
made = 54 773, .... 

Sharp beginning of graduated band, 

Faint band, 

Fainter band, .... 
Still fainter band, 


10 

h 

0-5 
1-0 

{ 2-0 

1-5 
0'9 

0-4 

0-3 


■ 

:i: 
■1: 

1 

- ! 


52 250 

52 538 

53 764 

54 242 

54 799 

55 358 

56 246 

56 304 

57 080 
57 570 

57 806 

58 212 


Carbon 

I Carbon 

[ Carbon 

[ 58 523 

Carbon 
Carbon 




54 799 

55 358 


Blue. 


Indigo. 




Violet. 


Violet Hydrogen, .... 

Part 5. 

Violet Hydrogen, .... 
Sharp beginning of faint violet band, 
Line, beginning of faint band, 
Probable thin line, 

Lavender Hydrogen, 

Possible faint haze extends thus far, 


4 

3 
1 
1 
0-3 

1-5 

o-i 


1 
1 

1 


58 521 

58 525 
58 930 
60 260 

60 860 

61 932 
63 090 


58 525 
61 932 


58 523 
61 932 


Lavender. 



GASEOUS SPECTRA IN VACUUM TUBES. 
IODINE ( = I). March 6 and 8, 1880. 



127 















6" condensed spark 


"j 














lines of Iodine. 


Left for Iodine 


Colour. 


Subject of Observation. 


Intensity 




Appear- 
ance. 


W.N. Place. 


Impurities. 






at low, as well 


Inten- 
sity. 


W.N. Place. 


as high, 
temperature. 




Part 1. 


















Clear line, .... 


1-5 


1 


36 490 




2 


37 020 




Ked. 


Do. do 


0'8 


1 


36 768 




2 


37 216 






Do. do 


2-0 


1 


37 334 




2 


37 590 






Do. do 

Part 2. 


0-5 


1 


37 792 




2 


37 967 








Strong line, . , , - . 


2 


1 


38 097 










Scarlet-Ked. 


Band of perhaps finer lines, 


1 




38 337 




2 


38 252 


I x 




Strong line, .... 


1-5 


1 


38 566 




2 


38 625 


x 




This is apparently Red Hydro- 
gen, 


J, 8 




38 711 


Hydrogen 






1 




Light-Red. 


Haze bands in a region of haze, 


1 & ) 
0-5 \ 

f 2-0 


» 1 


39 151 
39 408 

39 843 

40 049 




2 


39 113 


X 




or scarcely resolved linelets, 


"•"'1 




2 


40 069 


X 








0-2 


1 


40 203 














0-4 


, 


40 354 




2 


40 368 


X 






0-4 


1 


40 428 














0-4 


1 


40 502 














2-0 




40 586 




4 


40 594 


X 


Orange. 


All clear and rather coarse 
lines, .... 


l'O 

< 2-3 


1 
1 


40 734 
. 40 935 




4 


40 901 


X 






1 &) 
07 | 


'■ I 


41 235 




2 


41 172 








41 316 




2 


41 273 








3 


1 


41 466 




10 


41 428 


X 


1 




1 &) 
2-5 \ 


" t 


41 741 




2 


41 727 








41 854 




10 


41 824 


X 






I 1-0 
7 




42 051 
42 250 




2 


41 865 






Faint double line, . 


, 








Band of exquisitely graduated 


1-0 & 


Him j 


42 404 












thin lines, .... 


o-i 


42 579 












Brightest line yet, . 


3-5 


1 


42 709 




10 


42 646 


X 




Strong line, .... 


1-3 


1 


42 911 




2 


42 906 


X 




Line, 


0-9 


1 


43 088 




2 


43 131 


X 


Yellow. 


Part 3. 


















Clear line, .... 


0-6 


1 


43 236 




1 


43 299 


X 




Faint line, .... 


0-3 


i 


43 423 












Stronger line, 


0-6 


1 


43 540 












Do. do 


0-6 

f 2 


1 
i 


43 680 
43 867 




2 
4 


43 634 
43 867 


X 
X 








4 


1 


43 971 




10 


43 968 


X 






3 


1 


44 097 




10 


44 074 


X 






4 


1 


44 269 




10 


44 258 


X 






3 


1 


44 479 




10 


44 460 


X 






3 


1 


44 620 




10 


44 591 


X 






2 




44 723 




10 


44 695 


X 






1 




45 011 




2 


44 964 


X 






5 


1 


45 146 




10 


45 100 


X 






1 




45 245 




4 


45 196 


X 


Citron. 


All these lines clear and with 
black space between, but 
they are thick and coarse, . 


2-5 
0-2 
0-5 


i 


45 357 
45 442 
45 603 




4 
2 


45 300 
45 356 


X 






1 




45 738 




2 


45 700 


X 






1 


i 


45 971 




2 


45 931 


X 






1 




46 114 




4 


46 089 


X 






6-0 


1 I 


46 216 




10 


46 189 


X 






1-0 




46 335 


I 


2 
2 


46 232 
46 333 


X 






7-0 


1 


46 462 




10 


46 451 


X 






6-0 


1 


46 698 




8 


46 683 


X 






1-0 




46 827 




2 


46 847 


X 






I 7-0 


■' 


46 982 




10 


47 018 


X 



128 



PROFESSOR PIAZZI SMYTH ON 

IODIN E — continued. 



Colour. 



Subject of Observation. 



Intensity, 



Part 3 — continued. 



Green. 



Glaucous. 



All these lines are pretty clear 
and strong as such, but 
rather coarse, 



This brightest line of this 
Iodine, is wanting in the 
condensed spark spectra ; but 
is certainly here, having been 
independently identified by 
blow-pipe's Green Giant, 



A grand line (Int. 10, W.N 
= 49 437) of the condensed 
sparks' standard tables is 
wanting here, and should be 
re-observed, 



These Carbo-hydrogen refer- 
ences may be mere accidental 
coincidences, 



0-5 
3 

5&5 

1-0 
0-5 
1-0 
07 
2-0 
4-0 

1-5 
3-0 
0-3 
0-3 
2'0 
. 1-5 



HO 



\ 1 



All lines of various thicknesses, 



Band of thin lines, 



Faint Glaucous Hydrogen, 

Part 4. 

Trace of Glaucous Hydrogen, 



Lines again, but of coarse 
quality, .... 



0-3 
1-2 

0-3 
1-2 

3-0 

2 

0-5 

2 

2 

01 

01 

4 

2 

2 

2-5 

0-5 

3 
4 
3 



3 

3 

1-5 

1-5 

3 

0-3 

2 

3 

2 

0-4 

2 

1-5 



Appear- 
ance. 



II 



W.N. Place. 



Impurities. 



47 203 
47 300 
47 510 
47 582 
47 684 
47 829 

47 910 

48 016 
48 179 
48 380 

48 541 
48 676 
48 778 
48 868 

48 965 

49 063 



49 189 



49 316 



49 507 '| 



49 546 
49 639 

49 787 

49 863 

50 108 

50 307 
50 375 
50 475 
50 683 
50 762 
50 837 

50 893 

51 097 
51 212 
51 372 
51 525 
51 620 

51 726 

52 002 
52 224 



52 261 
52 386 
52 556 
52 627 
52 845 

52 978 

53 085 
53 337 
53 474 
53 616 
53 707 
53 911 



6" condensed spark 
lines of Iodine. 



Inten- 
sity. 



Carbo- 
Hydrogen 



( Carbo- 
| Hydrogen 



Carbo- 
Hydrogen 



Hydrogen 



Hydrogen 



10 
10 



10 



W.N. Place. 



47 237 
47 344 
47 574 
47 655 

47 799 

47 996 

48 270 
48 316 

48 520 
48 678 
48 761 



49 073 
49 167 
49 320 

49 437 



49 737 

49 783 

50 158 
50 327 
50 518 



50 901 

51 086 
51 209 
51 353 

51 605 

51 986 



52 340 
52 502 
52 565 
52 818 



Left for Iodine 
at low, as well 

as high 
temperature. 



? absent, or 

very weak, 

in 6" spark 

spectrum 



GASEOUS SPECTRA IN VACUUM TUBES. 

IODIN E — continued. 



129 



: 












6" condensed spark 

spectral lines of 

Iodine. 


Left for Iodine 


Colour. 


Subject of Observation. 


Intensity. 


Appear- 


W.N. Place. 


Impurities. 


at low, as well 








ance. 






Inten- 
sity. 


W.N. Place. 


as high, 
temperature. 




Part 4 — continued. 


0-5 
07 


i 


53 990 

54 152 










Glaucous. 




4 
4 
1 
3 


i "1 


54 270 
54 386 
54 480 
54 690 














2 


III 


54 777 




4 


54 788 


X 






2 


\ 


54 900 

55 038 














1-5 


\ 














1-5 


1 


55 168 














0-7 


I ■•! 


55 280 














07 


55 340 












Strong line preceded by a faint 
one, ..... 


I 2 


1 


55 452 












) 
1 


i 


55 665 










Blue. 




1-3 

1-3 

1-4 

1 

1 

1-5 

1-3 

2-0 


i 
i 
i 

i 

i 


55 851 

56 050 
56 238 
56 442 
56 560 

56 743 
"56 887 

57 023 
















1-5 




57 127 












Black space follows. 






1 










Indigo. 




2-0 


,;:| 


57 393 












f 0'3 


I 


57 501 














2-0 


1 


57 583 












All these iodine lines are bright 


1-5 


1 


57 727 












and clear, but thick and 


- 1-5 


1 


58 030 












coarse 


1*0 


1 


58 208 














1-3 


i 
1 


58 396 














L 1-2 


1 


58 501 












Faint indication, 


0-5 




58 560 


\ 








Part 5. 








i 58 530 








Trace of Violet Hydrogen, 


0-6 


l 


58 499 


) 








Violet. 


Line, 

A faint band, 

Hazy band of lines, 

Hazy line, .... 

Hazy band, .... 

Hazy band .... 


1-0 
0-3 

1-0 

07 
0-5 
1-0 
13 
1-0 


:i: 

i 


58 554 

58 786 

59 022 
59 268 
59 381 

59 718 

60 032 
60 258 
60 931 


1 












Broader haze band, 


1-2 




61 509 










Lavender. 


0'5 
0-2 




62 530 

63 104 


1 








End of Spectrum. 




This spectrum is most peculiar for its freedom from nearly allkno 


wn impurities of the other gases, and for consisting almost 


entirely of lines ; not however the sharpest order of lines, for thej 


t remind one more of straws than needles. They seem 


generally in position, very like the 6" condensed spark spectra line 


s ; especially if we allow that the one only strong case of 


divergence, 


viz. , that in the Green, is due to an ( 


nor in tl 


ie tabular 


spectrum re 


ferred to. 









VOL. XXX. PART I. 



U 



130 



PROFESSOR PIAZZI SMYTH ON 



MARSH-GAS. End-on Tube. October 8, 1879. 

= Methyl Hydride = Light Carhuretted Hydrogen = CH 4 . 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed for 
Marsh-Gas. 




Part 1. 












A few, and 


Red. 


Faint thick band, .... 


0-5 


» | 


36 964 ) 

37 402 \ 




Carbon 


only a few, 
of the ordi- 




Haze band, ..... 


0-8 


I 


37 616 

37 830 ) 

38 152 \ 




Carbon 


nary Carbo- 

hydrogen, 

but a grand 




Stronger hazy band, 


1-0 


it | 




Carbon 
















violet series 


Scarlet-Red. 


Part 2. 












of its own. 




Beginning of band going past Red 


|,0 




38 359 




Carbon ? 






Hydrogen's place, 














Red Hydrogen, .... 


5 


■ _ 


38 707 




Hydrogen 








End of previous faint baud, . 


0-5 




39 050 








Light-Red. 


Band cleft down middle, 


1-5 


■ i 


39 151 
39 532 


} Nitrogen 








Broad band, ..... 
Narrow band, .... 


2-0 
2-0 


-i 


39 630 

39 992 

40 151 


> Nitrogen 


Hydrogen 








Solid band, 


2-5 


!;, | 


40 262 
40 532 


! 


Carbon 






Narrow solid band, 


2'5 


, | 


40 630 
40 832 


f Nitrogen 
i j 








Strong line, ..... 


2 '5 


1 


40 984 




Hydrogen 




Orange. 


Band of just resolvable lines, 


2-0 


■ | 


41 085 
41 290 


[ Nitrogen 








Solid-like band, .... 


2-5 


il| 


41 390 
41 542 


I 


Hydrogen 






( begins hazily, 
Band < culminates in a Hue, . 
( ends hazily, 


1-5 
2 5 
1-5 


1 


41 638 

41 865 

42 018 


! 


J Hydrogen? 
( + Carbon? 








Very solid narrow band, 


3 


,j 


42 069 
42 246 


i 


Hydrogen 






Faint space of resolvable lines, 


1-5 












Band line, 


2 


| 


42 537 


? 








Faint space of resolvable lines, 


1-0 














Band line, ..... 


2 


| 


42 815 




Hydrogen 






Dark space, with faintest resolvable 


0-5 






i 








lines, 












Yellow. 


Stronger band line, 

Part 3. 

Line left off with in Part 2, . 
Fainter line, .... 
Dark space with faintest close lines 


2-5 

2-5 
1-5 

0-3 


1 

1 

1 


43 163 

43 157 
43 282 


' [ 43 160 


Hydrogen 






intervenes, .... 














Thick hazy line, .... 


2-0 




43 504 




Hydrogen 






Strongest line about, 


3-0 


1 " 


43 676 




Hydrogen 






Band of just resolvable lines, 


15 




43 723 

44 024 


| 


Carbon ? 








Probable double line, 


1-8 


II 


44 094 


) 


Hydrogen 






Strong thick green line, 


2-5 


I 


44 296 




Hydrogen 






Band of just resolvable lines begins, 


1-0 




44 456 


\ 








culminates in a thin bright 




/ \ 




1 








line, 


2'0 


> .::. < 


44 688 


} 


Hydrogen 






ends, ..... 


1-0 


) ( 


44 943 


) 






( 'itron. 


Narrow band of resolvable lines, . 


1-0 


- I 


45 012 
45 148 


) Carbo- 
) Hydrogen 




1 Carbo- 
( Hydrogen 




Citron band begins sharp, 
ends in haze, 


4 

1 


i 1 


45 250 
45 588 


i 


Carbon 






Then haze follows, 


0-5 














Hazy line, 


1-5 


1 


45 842 


I Carbo- 
\ Hydrogen 




I Carbo- 
( Hydrogen 




Haze intervenes, .... 
















Hazy line, 


1-5 


1 


46 195 


I Carbo- 

( Hydrogen 




I Carbo- 
j Hydrogen 




Haze intervenes, .... 















GASEOUS SPECTRA IN VACUUM TUBES. 

MAESH-GA S — continued. 



131 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Unclaimed for 
Marsh -Gas. 




Part 2 — continued. 














Citron. 


Hazy line, 

Haze intervenes, .... 


1-5 


1 


46 534 


? 




? 




Stronger line, .... 
Stronger line, .... 


2"0 


1 


46 876 

47 137 


? 




% 




2-0 


| 


2 




? 




Resolvable haze intervenes, . 
















Line, ...... 


L5 


1 


47 636 


Oxygen ? 








Resolvable haze intervenes, . 
















Thin line, ..... 


1-0 


1 


47 900 




Hydrogen 






Resolvable haze intervenes, . 
















Hazy band, ..... 


2-0 




48 263 


Nitrogen ? 








Very black space follows. 














Green. 


Grand Green band begins sharp, 


5 


1 ■■ i 


48 838 




Carbon 






Faint view of Blow-pipe Green Giant, 


1 


! fea - i 


49 181 


4 Carbo- 
( Hydrogen 








Fainter end of Green-band, . 


0-5 




49 410 










Resolvable haze band, . 


1-5 


m\ 


49 896 

50 356 










Hazy line, . . 


1-8 


50 682 




| Carbon? 

/ Hydrogen ? 






Hazy line, ..... 


1-0 


:-f: 


50 839 











Hazy line, ..... 


1-7 


=HH 


51 067 


Nitrogen ? 








Fainter hazy line, .... 


1-0 




51 275 


i 




? 




Farrow band, 


2 


."1 


51 463 
51 559 


I 


Hydrogen 






Very faint hazy line, 


0-5 


51 726 


Nitrogen ? 








Glaucous Hydrogen, 


4 


1 


52 256 


[ 52 258 


Hydrogen 






Part 4. 












Glaucous. 


Glaucous Hydrogen, 


4 


1 


52 260 








Blue hand \ Be S iDS snfu 'Ply. 
Utue band j Endg faintlVj 


3 

0-8 


1 !ln„ | 


52 544 

53 112 


1 


Carbon 






Faint band 1 Begins sharp, . 
*amt band | Ends weakj 


1-0 

0-3 

( 0-5 


1 *■ 1 


53 803 

53 993 

54 182 


? 




? 




Hazy lines, 


I 0-5 
( 0-5 




54 420 
54 628 











Stronger hazy line, 
Faintest hazy line, 


1-0 
0-2 


ill 3 


54 787 

55 094 




Hydrogen 






Band \ Be S ins sharply, 
i5anfl | Ends weakly, . 


1-5 
0-5 


! -• t 


55 326 
55 740 


1 


Hydrogen 




Blue. 


Faint haze, . . . . . 


03 




55 954 








Faint haze, 


0-3 




56 054 










Strong haze band, begins, 
ends, 


2-5 

1-0 

( 1-5 


'I 


56 200 

56 732 

57 488 


} 


Carbon 








? 


Tndigo. 


Haze bands, ..... 
A haze band begins, including 


1-0 
( 0-7 

l'O 


) '' ( 


57 897 

58 147 

58 342 






i 






Violet Hydrogen, 


,=! 










Violet Hydrogen, .... 


1-5 


) I 


58 615 


) 


Hydrogen 






Part 5. 








I 58 514 








Violet Hydrogen, .... 


2-5 


I 


58 512 


) 






Violet. 


Faint line, ..... 


1 


i 


58 927 


| Carbo- 
( Hydrogen 




Carbo- 
Hydrogen 




Line beginning a faint band, . 
End of that band, .... 


2-0 
0-3 


I ' I 


59 474 
59 700 


| Nitrogen ? 








Very black space follows, 

















Leader of peculiar Marsh-gas series, 


3 


I 


60 243 






60 243 




Second, a hazy line, 


2-0 


II 


60 534 






60 534 




Third line, hazy, .... 


1-5 

r l-o 

1 0-8 




60 704 

60 948 

61 161 






60 704 
60 948 












61 161 




Remaining lines of "Marsh-gas 


J 0-7 
' 0-5 




61 355 






61 355 


Lavender. 


series" weaker, hazier and broader, 




61 563 






61 563 




, 


0-3 




61 865 






61 865 






I 0-4 




62 139 






62 139 



1 32 



PROFESSOR PTAZZI SMYTH ON 
MAESH-GA S— continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. ' 

| 


Impurities. 


Constituents 
dissociated. 


Unclaimed for 
Marsh-Gas. 


I. . nder. 


Part 5 — continued. 
Solitary hazy line, 
Very faint haze band, . 

Faint but sharp beginning of a band, 
looking suspiciously like a reflec- 
tion, ...... 


1-0 
0-3 

( 0-5 


. 


62 548 

63 791 

65 610 




Carbon ? 


62 548 

63 791 

65 610 


Gray. 

1 



End of Spectrum. 

The beautiful violet series of lines and bands, beginning at 60 243, and continuing to 62 000 nearly, in decreasing 
intensity was discovered by Prof. A. S. Herschel in these End-on tubes in 1879, and confirmed therein by myself; with 
the result of finding traces of them in some other hydro-carbons, as Alcohol ; but always far weaker than in Marsh-Gas. 

In hardly any Hydro-Carbon is there so little of the Marsh-Gas series to be seen as in defiant Gas ; and yet that is so 
near in chemical constitution to Marsh-Gas, which has the series so splendidly, that it must belong to that gas ; and it is 
certainly the spectrum of a compound, and not an elemental, gas, because it vanishes with greater intensity of spark and 
dissociating power. 



NITROGEN. 



End-on Tube. Observed July 16, 1879. 

Symbol=N. 



Colour. 



Crimson- 
Red. 



Red. 



Scarlet-Red. 



Light-Red. 



Orange. 



Yellow. 



Subject of Observation. 



Part 1. 

Very faint hazy line, 
Hazy line, 
Haze band, . 
Haze band, . 

Do. do. . 

Do. do. 

Do. do. 

Do. do. . 
Red Hydrogen, 



Part 2. 

Red Hydrogen, 
Faint band, . 

Hazy band, . 

Do. do. . 

Do. do. . 

Do. do. . 

Hazy line, 
Do. do. . 

The Oxygen line, 
Faint hazy band, 



Hazy band line in haze, 
Do. do. 



Intensity. 



0-5 
1 


1-0 
1-5 
1-5 
1-5 




0-5 

2 

2 
2 
2 

1 
1 

3 
1 

3 
3 



Appear- 
ance. 



W.N. 


Place. 


1 

32 555 


33 


975 


34 


973 


36 


189 


37 075 


37 


555 


37 


996 


38 


430 


38 


705 


38 


708 


38 


988 


39 


260 


39 


588 


39 


692 


40 


024 


40 


150 


40 


514 


40 


628 


40 


930 

1 


40 997 


41 


130 


41 


255 


41 


460 


41 


736 


41 


818 


42 


005 • 


42 


120 


42 


320 



■»-■ SSS 1 



Oxygen 



Hydrogen 
Hydrogen 



Hydrogen ? 
Hydrogen ? 

Oxygen 



Carbon and 
Nitrogen 



Unclaimed 
features left 
for Nitrogen. 



33 975 

34 973 

36 189 

37 075 
37 555 

37 996 

38 430 



38 988 

39 260 
39 588 

39 692 

40 024 
40 150 
40 514 
40 628 
40 930 



41 


460 


41 


736 


41 


818 


42 005 


42 


120 


12 


320 



GASEOUS SPECTEA IN VACUUM TUBES. 
NITROGEN. End-on Tube — continued. 



133 



Colour. 



Subject of Observation. 



Yellow. 



Intensity. 



Part 3 — continued. 

Hazy band line in haze, 

Do. do. 

(reading altered from 42 

Do. do. 



746). 



Do. 



do. 



Yellow line in haze, 

Part 3. 

Yellow line in haze, 

Haze band, . 
Faint haze band, . 
Broad line, . 



Region of broad dark faint bands 
generally, 

First of several bright haze bands 
in haze, ..... 

Bright haze band in haze, 



Do. 



do. 



Haze band in haze, 
Narrow band in haze, 

Do. do. . 

Do. do. . 

Band in haze, 



Faint bands, and then comes a 
strong sharp-edged band still in 
haze, ..... 

Narrow band in haze, . 

Band in haze, 

Congeries of faint lines in haze, 

Grand Green band begins sharp, 
It is composed thus, viz. — 

Strong hazy band, 
Less strong, 
Very faint, 

Decreasing nebulous haze ; no Green 
Giant of Blow-pipe appears, 

/Blow-pipe's Green Giant's place by 
V a separate Alcohol tube is . 

Part 4. 

A fainter Green band begins, 

Glaucous band 

Second glaucous band, 



Glaucous Hydrogen. 



3 
3 

2 

3 

0-5 

4 
4 



Appear- 
ance. 



3 
3 
2 

5 

5 
3 
1 

07 to 
0-1 



3 
3 

2 

10 



W.N. Place. 



Impurities 



42 402 
42 588 

42 719 

42 946 

43 041 
43 296 
43 364 
43 582 

43 673 



43 687 

43 783 

44 009 
44 109 
44 239 
44 290 
44 423 

44 500 

45 224 

45 241 
45 496 
45 634 
45 884 

45 995 

46 230 
46 379 
46 547 
46 733 
46 867 

46 993 

47 210 

47 488 
47 716 

47 800 

47 920 

48 120 
48 316 
48 442 
48 750 

48 822 

48 802 
48 929 

48 967 

49 062 
49 142 
49 242 

49 270 

50 070 

49 172 ' 



50 068 

50 962 

51 577 

52 209 



Hydrogen 
43 680 



Hydrogen 



Carbon and 
Nitrogen 



Oxygen ' 



Nitrogen 
and Oxygen 



Constituents 
dissociated. 



Carbon 
and 

Nitrogen 



Hydrogen 



Unclaimed 
features left 
for Nitrogen. 



42 402 
42 588 

42 719 

42 946 

43 041 
43 294 
43 364 
43 582 



43 


783 


44 


009 


44 


109 


44 


239 


44 


500 


45 


224 


45 


241 


45 


496 


45 


634 


45 


884 


45 


995 


46 


230 


46 


379 


46 


547 


46 


867 


46 


993 


47 


210 


47 


488 


47 


716 


47 


800 


47 


920 


48 


120 


48 


316 


48 


442 


48 


750 


48 


802 


48 


929 


48 


967 


49 


062 


49 


142 


49 


242 


49 


270 


50 


070 



50 068 

50 962 

51 577 



134 



PROFESSOR PIAZZI SMYTH ON 

NITROGEN. End-on Tube— continued. 



Colour. 



Glaucous. 



Blui . 



Subject of Observation. 



Part 4 — continued. 
Glaucous blue band, 
Another band close upon it, . 



A doubled band 

Bright bars at beginning, then 
shade, and finally blackness be- 
fore the next bright bar, 



Intensity. 



Appear- 
ance. 



W.N. I'laee. 



52 488 

52 719 

53 780 

54 435 

54 634 

55 515 



Impurities. 



Carbon 



1 



Constituents 
dissociated. 



N.D. — In the Green the last of each band's space was hazy, not dark. 

55 565 



Indigo. 



Violet. 



Lavender. 



Part 5. 

Beginning of band, repeated, 
Do. do. 

Do. do. 

Do. do. 

Do. do. 



Violet Hydrogen, centre of line, 



Sharp beginning of band, 
Haze band, . 
Sharp beginning of band, 
Do. do. 



Do. 
Do. 
Do. 



do. 
do. 
do. 



Very faint and uncertain, 



4 




1 




4 


::::. 


4 




3 




4 


1 


0-5 




4 




1 




4 




3 




2-5 




2-5 




2 




0-5 





55 615 

56 310 

56 555 

57 470 

58 256 

58 489 

59 351 

59 444 

60 247 

60 458 

61 320 

62 040 

62 636 

63 638 

64 605 



Carbon 



Hydrogen 



Carbon ? 



Unclaimed 

features left for 

Nitrogen. 



52 719 

53 780 

54 435 

54 634 

55 565 



55 565 

56 555 

57 470 

58 256 



59 444 

60 458 

61 320 

62 040 

62 636 

63 638 

64 605 



One more band still, but it looks like a glare-reflection of Violet Hydrogen and its close preceding bands, and reads 
= W.N. 65 517, but it may be the band seen following the Marsh-gas series and in that case probably Carbon. 

Alter this, darkness, and the end of the Spectrum. 

This spectrum, looked on by M. Plucker as the Spectrum of pure Nitrogen, but the band, or compound-line, or low 
temperature, form of the same, — is stated by M. Thalen to be, on the contrary, the spectrum of the Compound of Nitrogen 
and Oxygen (bi-oxide of Azote) ; and if asked whence the oxygen for the Nitrogen of the tube to compound with, — he would 
say, from the two dissociated elements of watery vapour lurking in the tube, for see how large an uncombined amount of 
the other element, hydrogen, there is present. 

But if our "Water" and " Salt- Water " tubes show little or no dissociated elements of water, — we conclude that our weak 
sparks cannot dissociate accidental moisture of water either ; and that there are in this tube, pure nitrogen giving a band 
spectrum, a large amount of hydrogen impurity, a small quantity of carbon, perhaps carbo-hydrogen, impurity, but only a 
trace of oxygen impurity. 

Whence then has come so large an amount of Hydrogen by itself? 

It may have been liberated by the action of the spark in the vacuous tube out of the electrode wires and their "occluded" 
stores ; or out of the material of the glass itself. Or again it may be another example of the cases mentioned in the Intro- 
duction (pp. 101 to 103), of nitrogen, when acted on in a high state of rarefaction by the electric spark, changing into, or 
giving out, hydrogen. 



GASEOUS SPECTffcA IN VACUUM TUBES. 



135 



NITROUS OXIDE. Laughing-Gas. 

August 19, 1879. 



End-on Tube. 



=N 2 0. 



Colour. 



Crimson- 
Red. 



Red. 



Scarlet-Red. 



Light-Red. 



Orange. 



Yellow. 



Subject of Observation. 



Citron. 



Part 1. 

Hazy maximum line of a band of 

haze 

Another like it, 
Others similar as far as, 

Haze band, ..... 

Do. do 

Do. do. . . . . . 



Intensity. 



0'2 

0-2 
0-3 



1-0 

ro 

1-5 



Appear- 
ance. 



{N.B. — One more such band ; and then comes Red Hydrogen.) 



Part 2. 

Haze band, 



Red Hydrogen (not very strong 
only an impurity), 

Haze band, .... 

Do. do 

Do. do 

Hard-edged haze band, . 



Do. 



do. 



A distinct line (Ox.?) on Neb 
haze, ..... 

Fainter line in haze region, . 

Sudden beginning of hard, bright 
haze bands generally with faint, 
dark, double line down middle 



Do. 



do. 



Do. 


do 


Do. 


do 


Part 3. 




Hard haze band, 




Do. do. 




Do. do. 





Last sharp edge of broad haze band, 

Several thin lines, 

Sharp edge of brilliant band, . 
Region of excessively thin close 

lines, . 

Line among fainter ones, 

Do. do. 

Do. do. 



15 

5 



3 

2-5 
3 

1-5 
1-0 
4 

2 

2 
2 
2 



W.N. Place. 



33 560 



33 


860 


36 


920 


36 


970 


37 


224 


37 


426 


37 


758 


37 


856 


38 


234 



Impurities. 



Constituents 
dissociated. 



Nitrogen 



Nitrogen 
Nitrogen 
Nitrogen 



38 285 
38 340 

38 701 

38 894 

39 110 
39 262 
39 532 

39 674 

40 046 

40 130 
40 442 
40 578 

40 890 

41 234 



41 470 

41 714 

42 130 

42 162 

42 470 
42 600 

42 912 

43 006 
43 360 



43 342 
43 664 

43 756 

44 032 
44 124 
44 400 

44 799 

44 845 

45 167 
45 245 

45 613 

46 081 
46 429 
46 561 
46 725 



Hydrogen 



Hydrogen 

Carbon and 
Nitrogen 



Carbon 



Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 
Oxygen 

Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 

Nitrogen 

Nitrogen 
Nitrogen 
Oxygen ? 



Left unclaimed 

for 
Nitrous Oxide. 



None of any 
importance. 



136 



PROFESSOR PIAZZI SMYTH ON 
NITROUS OXID E— continued. 



Colour. 



Citron. 



Subject of Observation. 



Intensity. 



Appear- 
ance. 



Green. 



Glaucous. 



Blue. 



Indigo. 



Violet. 



Lavender. 



Part 3 — continued. 

Line among fainter ones, 

Broad faint resolvable band, . 

Bright line on faint haze, 
Resolvable fainter haze intervenes. 

Haze band ended by a line, . 

Bright beginning of a grand band, 

A line just visible near middle of 
that band, .... 

Part 4. 

From Green Giant's pkee, there is 

faint haze to this line, 
Haze intervenes, then this line, 
Fainter haze, and then thisline-band, 

Do. do. do. 

Faint haze follows. 
Glaucous Hydrogen, not very bright, 

Beginning of a flat band, 

Sharp beginning of a graduated 

band, 

Sharp beginning of a band, . 

Beginning of a fiat band, 

Sharp beginning of a graduated 

band, ..... 

Sharp beginning of a graduated 

single band, .... 

1 Beginning of a flat band, 
3rd case < Sharp beginning of a 

( graduated band, . 
Sharp beginning of a graduated 

band, 

Do. do. 

Violet Hydrogen cuts in on the 
above, ..... 

Part 5. 

Violet Hydrogen, .... 

Line or rib, beginning graduated 
band, ..... 

Very faint line, .... 

Rib beginning band, 

Do. do 



Do. 
Do. 
Do. 



do. 
do. 
do. 



Problematical, 



1-5 
2 



1-5 

1 

1 
1 



3 
3 

2 '5 

2 

2 
3 

2 
2'5 

2 
1-5 

2 



0-3 

2 

1-5 

1-3 
0-9 

0-5 

o-i 



W.N. Place. 



46 889 



iiiii j 


47 010 


47 440 


i 


47 850 


,;lj 


48 130 


48 450 




48 850 



49 170 



50 124 

50 437 

51 030 

51 632 

52 230 

52 505 

52 752 

53 794 

54 420 

54 640 

55 484 

56 280 

56 610 

57 560 

58 340 

58 528 



58 519 



59 530 

60 278 

60 500 

61 350 

62 102 

62 672 

63 706 

64 560 



Impurities. 



Hydrogen \ 



Carbon 

( Carbo- 
( Hydrogen 



Hydrogen 
Carbon 



Carbon 



■"l Hydrogen 
I- 58 524 



Carbon 



Constituents 
dissociated. 



Left unclaimed 

for 
Nitrous Oxide. 



Nitrogen 

Nitrogen 
Oxygen ? 

Nitrogen 



Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 



Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 

Nitrogen 

Nitrogen 
Nitrogeu 



Nitrogen 

Nitrogen 
Nitrogen 

Nitrogen 
Nitrogen 
Nitrogen 

Nitrogen 



End of Spectrum. 

This spectrum is very like that of Nitrogen ; but with less than half the amount of Hydrogen impurity, and a similarly 
small proportion of Oxygen, though that should be present as a constituent dissociated, at the same time that the Nitrogen 
was freed; but Oxygen is undoubtedly a bad illuminator. There are also large traces of Carbon, and smaller of Carbo- 
hydrogen, impurities. 



GASEOUS SPECTRA IN VACUUM TUBES. 



137 



OLEFIANT GAS. End-on Tube. October 9, 1879. 

= Ethylene = Heavy Carburetted Hydrogen =C 2 H 4 . 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 

II 


Constituents 
dissociated. 


Left unclaimed 

for 

Olefiant Gas. 


Red. 


Part 1. 

Faint haze band, .... 
Do. do 

Broad haze band, .... 
Do. do. ' . 

Part 2. 

Red line beginning a band, . 
Red Hydrogen breaks in on that 
band, 


0-3 
0-4 

1-0 

1-0 

1-5 


- 1 

1:.-. 
■ 


35 846 

36 721 

36 895 

37 372 

37 642 

38 020 

38 397 
38 715 


i 
1 


Carbon 

Carbon 
Carbon 

Carbon 
Hydrogen 


None 

distinct from 

Carbo- 

hydrogen ; 

but that, 

Olefiant gas 

claims the 

utmost part 

in, of all the 

gases here 

tried. 


Scarlet-Red. 




Light-Red. 


f Bundle of haze bands begins, 

1 
-{ Brighter lines therein, 

L End of the bundle of haze, . 
Narrow band or broad line, . 

Band almost resolvable into lines, . 

A very thin line intervenes, . 


1-0 

i 1<5 
1-5 

( 1-5 
0-5 

2 

2-5 
1-0 


1 

i 


39 055 
39 277 
39 452 
39 678 

39 958 

40 138 
40 288 
40 561 
40 591 


Oxygen ? 

i 


Hydrogen 
Hydrogen 
Hydrogen 
Hydrogen 
Carbon? and 
Hydrogen ? 
Hydrogen 






Orange. 


Resolvable band, .... 

Very black but narrow space follows. 
Broad line, ..... 


2-5 
2-5 


1 


40 617 

40 875 

41 003 


) 
\ 

1 


Hydrogen 
Hydrogen 






Resolvable band, .... 


2-0 


. 


41 096 
41 308 


> Oxygen ? 








A brighter resolvable band, . 


2-5 


:::: j 


41 373 
41 560 


j 


Hydrogen 






Very black line or space intervenes. 
Resolvable band begins, 

Bright line therein, 
ends, . 


1-0 
3-5 

1-0 


! -1 


41 663 

41 748 

42 044 


1 1 


Hydrogen ? 
Carbon ? 








Narrow hard band, 


3 


!! I 


42 133 

42 288 


I 


Hydrogen 




Yellow. 


After a dark space, a resolvable 
band, ..... 

Dark space with very thin lines, . 

Thin double line ? . 

Dark space with very thin bright 
lines, ..... 

Another bright line, 


| 2-0 

0-5 
2 

| 0-5 

2 


» 


42 430 
42 612 

42 799 

43 157 




Hydrogen 






Part 3. 

Bright line left off at before, . 
End of the attendant resolvable 
band, ..... 
Thick faint line, .... 
Very bright line perhaps double, . 


2 

|,» 

1-5 

3 '5 


J 

1 
1 


43 177 

43 367 

43 575 
43 746 


( 43 167 


Hydrogen 

Hydrogen 
Hydrogen 






Faint band of just resolvable lines, 
Narrow band, .... 


1-0 

2-0 


■ i 
■1 


43 783 

44 166 
44 273 
44 398 


i 

s 


Hydrogen ? 
Hydrogen 






Citron. 


( Broad beginning of another band, 
< Thin bright line, 
( End of that band, 


1 
2 
0-3 


M 


44 490 
44 721 
44 944 


1 


Hydrogen 






Strong line (Citron line 1), 
Sharp beginning of band, 


3 

4 


1 


45 099 
45 316 


\ Carbo- 
j Hydrogen 


Carbon 


Carbo- 
Hydrogen 




Supposed Citron line 2, 


2 


1 


45 521 


I Carbo- 
/ Hydrogen 




Carbo- 
Hydrogen 




After haze, another line, 


2 


1 


45 901 


i Carbo- 
| Hydrogen 




Carbo- 
Hydrogen 




After more haze, another line, 

1 


1-5 


' 


46 201 


j Carbo- 
j Hydrogen 




Carbo- 
Hydrogen 



VOL. XXX. PART I. 



X 



138 



PROFESSOR PIAZZI SMYTH ON 

OLEFIANT GA S— continued. 



Colour. 


Subject of Observation. 


Intensity. 


1 

Appear- 
ance. 


W.N. Place. I 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 

for 
Olefiant Gas. 




Part 3 — continued. 














Citron. 


After more haze, another line, 
Sharp concluding edge of a band, . 


1-0 

1-5 

r 2 


1 


46 495 

46 901 

47 143 


1 Carbo- 

( Hydrogen? 


Carbon ? 


Carbo- 
Hydrogen 




Hazy lines 


) 1>5 

) 1-0 

( 0-7 


III 
:i: 


47 427 
47 651 
47 914 


Oxygen ? 


Hydrogen ? 






Narrow baud ends sharply, . 


2 


4 \ 


48 120 

48 298 










Dark space follows. 














Green. 


Sharp beginning of Green Band, 


5 




48 844 ! 




Carbon 






Green Giant of Carbo-Hydrogen 


1' 


l 


49 178 


1 Carbo- 




Carbo- 




and Blow-Pipe, .... 


j Hydrogen 




Hydrogen 




Decreasing haze of the band follows, 


2 














Second line of Carbo-hydrogen, 


4 


I 


49 548 


I Carbo- 
( Hydrogen 




Carbo- 
Hydrogen 




Decreasing haze follows, 


0-5 














Bundle of haze, resolvable, begins, 


1-0 


) :: ' I 


49 874 


) 








maximum, 


2-0 


III!! 


50 179 




Hydrogen ? 






ends, 


0-5 


) \ 


50 454 










Narrow haze band, 


I? 


\ 1 


50 638 

50 847 


1 


Carbon 






Band, 


IB 




50 996 

51 304 












Possible double line, 


\\ 


MM 


51 457 
51 536 










Glaucous Hydrogen, 


4 


1 


52 263 


j 


Hydrogen 






Part 4. 








[ 52 262 






Glaucous. 


Glaucous Hydrogen, 


4 


1 


52 262 


1 


Hydrogen 






Blue band begins sharp, 


3 


!■■■ \ 


52 544 


1 


Carbon 






ends weak, . 


0-5 


ji=::. | 


53 056 


| 






Very thin faint line, 


0-5 




53 258 










(Probable blue band line of Carbo- 


}« 


1 


53 660 


) Carbo- 




Carbo- 




Hydrogen) Line, 


\ Hydrogen 




Hydrogen 




Narrow band, .... 


1-0 




53 891 


i Carbo- 
\ Hydrogen 




Carbo- 
Hydrogen 




( Begins, . 


0-5 


) ( 


54 129 


) 








Broad haze band \ Maximum, 


1-2 


\m 


54 258 


f 


Carbon 






( Ends, 


0-5 




54 468 


\ 








Sharp line, ..... 


2-0 


\ 


54 790 




Hydrogen 






Blue. 


Band begins, very sharply, . 


1-5 


\ j -- ! 


55 391 




Hydrogen 






ends faintly, 


0-3 




55 607 










Grand violet band begins, 


25 
0'3 


1 «'■ 


56 266 

56 776 

57 137 


\ 


Carbon 




Indigo. 


Faint haze band, .... 


0-5 


) / 


i 








Faint haze band, .... 


0-5 




57 512 










Second violet band, begins, . 
ends, 


2 
0-5 


\ -\ 


57 807 

58 259 


I 


Carbon 








Violet Hydrogen, .... 


1-5 


i 


58 539 




Hydrogen 




Violet. 


Part 5. 








[ 58 540 








Violet Hydrogen 


1-5 


i 


58 542 


) 








A band beginning with a sharp line, 


1-8 


it j 


58 911 


j Carbo- 


[ Carbon 


| Gar-bo- 




ends, .... 


0-4 


59 117 


^ Hydrogen 


( Hydrogen 






End of broad haze extending from 
last, ...... 


I 0-4 




61 207 








Lavender. 


Black space follows. 
















Gray band begins sharply, 


1-5 


is, i 


61 567 


1 


flflvhnn 






ends weakly, 


0-2 


62 213 


'—ill tJKJll 






End of all haze light, 


o-i 




62 972 









End of Spectrum. 

In this spectrum there is merely a trace of impurities absolute of Nitrogen and Oxygen. The Constituents Carbon and 
Hydrogen are marked, but not extravagantly ; partly perhaps because they also appear together as Carbo-hydrogen lines rather 
signally. No Marsh-gas series appears here ; but Marsh-gas had not the Carbo-hydrogen lines by any means so strongly. 



GASEOUS SPECTRA IN VACUUM TUBES. 



139 



OXYGEN. End-on Tube. Spark = 0"-8. August 11, 1879. 

Symbol=0 . 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance. 


W.N. Place. 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 
for Oxygen. 


Crimson- 
Red. 


Part I. 

Faint but thin sharp line, 
Faint beginning of haze, 
Sharp ending of same haze, . 


0-5 

j,. 


t 


32 600 

33 150 

34 240 


I • 




32 600 


Red. 




Sharp line (Ox.?), 

Haze, 

Part 2. 
Red Hydrogen, brilliant but blurred, 


1-5 
0-7 

10 
3 


1 

, j 

u 
1, 


36 270 

36 636 

37 950 

38 705 

39 360 


i ■ 

Hydrogen 




36 270 
39 360 


Scarlet- 
Red. 




ULI UlJi; IlllL- V V.v. Jj i 


■ • 


Light- 
Red. 


Broad haze band, . 


• 


1 


111! j 


39 600 
39 964 


I ! 








Narrow haze band, 
Resolvable haze band, 


■ 


1 
2 


J | 


40 126 

40 308 
40 610 


Hydrogen 
) Carbon ? & 
\ Hydrogen ? 










Resolvable haze band, 




2 


in,! 


40 662 

40 870 


1 Hydrogen? 






Orange. 


Hazy line, 
Infinitely thin line, 


• 


1-5 
0-2 


in 

i 


40 954 

41 110 


Hydrogen 
Hydrogen 








Grand (Ox ?) line, . 




4 


1 


41 266 






41 266 




Infinitely thin line, 
Hard edged band, . 

T .1T1P Wifil n A 7P Tiqp Irf/iTH 


„,i 


0'2 
1-5 
1-5 

1-5 


% ! 


41 662 
41 766 

41 922 

42 026 

42 120 
42 220 


) Carbon ? 
\ Hydrogen? 

1 








I *• lit! VV 11 1 i lli t/ijL- > h U. ls_m 1.' { i 1 1 i . . . 

Double line in haze, 


1; 

mi j 


> Hydrogen 








Faint but resolvable band, 


1-0 


»| 


42 310 
42 552 


i ! 






Yellow. 


A single line, .... 
Line, perhaps double, . 
Band of lines, .... 
Thick line (a yellow line past 
Sod. a), 

Part 3. 

The last yellow line (not Sod. a) . 

Several thin sharp, but faint lines, 

Strongest line yet, 


1-7 
1-0 
07 

j,, 

1-5 
( 0-7 
{ 07 
( 0-7 

2 


i ' 

i 

III! 
1 

1 

1! 
1 
| 
1 


42 640 

42 809 

42 964 

43 152 

43 132 
43 250 
43 394 
43 530 
43 683 


i 

Hydrogen 
Hydrogen ? 

1 Hydrogen 
f 43 142 

J 
Hydrogen 

? 

Hydrogen 
Hydrogen 




42 640 




Lines in haze, .... 
Broad line, 


11 

1-5 


:i; 

" r i" 


43 806 

43 990 

44 128 
44 312 


Hydrogen 
Hydrogen 










Line in haze, .... 
Faint line in haze, 


1 

0-7 


■ii 


44 532 
44 647 


Hydrogen 
Hydrogen 






Citron. 


Band just resolvable into lines, 


0-5 


-i 


44 794 

45 166 


i ! 








Sharp beginning of a brilliant band, 

Thin lines in the long tail of above 
band, 

Brilliant clean line (Ox ?), 


4 
( 1 
) 0-5 
) 1 
( 1 

3 


i 

i 


45 248 
45 754 

45 900 

46 108 
46 520 
46 724 


Carbon 
Hydrogen ? 




46 109 
46 724 




Green. 


Faint, but edged band, . 


0-5 


.i 


47 090 

47 458 









140 



PROFESSOR PIAZZI SMYTH ON 

OXYGEN. End-on Tube — continued. 



Colour. 


Subject of Observation. 


Intensity. 


Appear- 
ance; 


W.N. Place. I 


Impurities. 


Constituents 
dissociated. 


Left unclaimed 
lor Oxygen. 




Part 3 — continued. 

Another sharp brilliant line, . 


4 




47 676 






47 659 




Ultra faint haze band, . 


0-2 


I 


48 130 
48 450 










Very thin faint line, 


o-i 


48 710 








Green. 


Sharp beginning of Green band, 
Thin line iu that band, . 

Part 4. 
The thin line last alluded to, 


5 
1 

1 


1 


48 840 

49 360 

49 370 


Carbon 

r Hydrogen ? 
( 49 365 








Faint, semi-resolvable haze, . 


0-5 


■ ! 


49 896 

50 452 


| Hydrogen? 








Faint but hard edged band, . 

Faint line, 

Fainter haze intervenes, 

Faint line, ..... 


0-5 

1 

0-3 

1 


1 

i 


50 588 

50 858 

51 120 
51 490 


) Carbon ? 
( Hydrogen? 

Hydrogen 








Glaucous. 


Glaucous Hydrogen, excessively 
bright, ..... 

Glaucous Hydrogen, repeated after 
improving focus, 


!• 


■ 
■ 


52 246 
52 250 


Hydrogen 
Hydrogen 








Blue band 

Faint hazy line, in fainter haze, 

Do. do. 

Do. do. 
End of broad faint haze, 

Sharp beginning of faint band, 
Sharp beginning of flat band, 
Strong line ends it. (Ox.) . 


3 

1 

0-7 

07 

0-3 

1-5 
0-5 

2 




52 514 
54 020 
54 286 

54 820 

55 990 

56 286 

57 804 

58 160 


Carbon 
Carbon ? 

Carbon 
Carbon 




58 156 


Blue. 


Indigo. 




Violet Hydrogen, .... 


3 


1 


58 525 


\ Hydrogen 






Violet. 


Part 5. 
Violet Hydrogen, .... 
Beginning of faint haze, 

End of same, .... 
Gray band, 


3 
3 

o-i 

0-3 


1 

\ 

\ 

■■■■■■ I 


58 540 
58 930 

61 160 
61 658 
61 840 


V 58 532 

) Hydrogen 
Carbon ? 






Lavender. 




Gray line 


1-0 


■ 


61 904 


Hydrogen 







End of Spectrum. 

No Nitrogen appears here. 

But much Carbon and Hydrogen impurity, as well as the proper Oxygen ; the latter too in far greater force than in any 
other tube. 

The two first red lines re-observed on April 7th as 32 568 and 36 134. 

Of the other and stronger Oxygen lines, two certainly, and probably four, are very close, exceedingly close and beautifully 
sharp, doubles; of which I hope to give a further account on a future occasion, after completing some arrangements now in 
progress for increasing both the dispersion and the magnifying power of my present spectroscope. 



GASEOUS SPECTRA IN VACUUM TUBES. 141 

OZONE. End-on Tube. Spark = 0" -85. August 4, 1879. 

Symbol =0 3 . 

This spectrum has much Carbon, also Hydrogen, impurity ; but otherwise only shows pure Oxygen lines 
like an Oxygen tube, but not quite so brightly. Its tables therefore have been dispensed with for economy 
of printing. 



SALT WATER. End-on Tube. October 3, 1879. 

H 2 + Na. 

This tube yielded plenty of Hydrogen lines, but none of Oxygen, none of Na, or common Salt, and no Solar 
"rain-band" lines. Its numerical tables have therefore been suppressed. 



WATER. End-on Tube. September 29, 1879. 

Compound =H 2 . 

This tube showed strong Hydrogen, but no Oxygen and no Solar "rain-band" lines. Its numerical tables 

have therefore been suppressed. 



It seems probable now, that "rain-band" lines, or the lines and bands of Watery vapour as seen in the 
Solar spectrum, are not emission lines reversed, but pure absorption effects. I was indeed told some years 
ago by a great spectroscopist, that any induction spark in ordinary moist air, on a drizzly day, would show 
the water-vapour lines of the solar telluric spectrum as bright lines : but the only printed observation he 
has furnished me with, refers to a band of lines that can be photographed far away in the ultra-violet, non- 
visible, region of the spectrum : and all the experiments I have tried myself with such induction sparks as 
I have hitherto been able to command (very small and poor unfortunately) have never shown me anything 
bright connected with water or steam in free air, or small tubes, in the spectrum places, or with the charac- 
ters of Solar little a and its preliminary band of lines, the lines grouping about C, or the lines forming the 
chief practical rain-band for Meteorology, viz., the band on the red side of D. 

The artificial production, and final proof, of these wdl probably never be obtained, until very long tubes 
both of water and steam (such as only the nation, not an individual, could afford to set up) are used to 
intercept a strong light of known spectral quality. In so far, as in M. Janssen's celebrated experiment 
with the high-pressure steam-tube, but whose observations have never been clearly or numerically published. 

Equally too do all the known gases and their fluids, both elemental and compound, require to be experi- 
mented upon, as to their sheer absorption effects and nothing else ; while I have had some recent proofs 
with a good spectroscope, that the lines forming the foundations of some usually hazy absorption bands, are 
often as sharp, distinct and characteristically grouped, as anything ever exhibited by emission lines, whether 
direct, i.e. as bright lines, like all those which are noted in this collection of gaseous spectra (though the 
symbols for them are, for practical printing, made black on white), — or reversed, i.e. as black lines on a 
continuous bright spectrum, as with the ordinary " Fraunhofer " so-called lines in the spectrum of the Sun. 



142 PROFESSOR PIAZZI SMYTH ON 



APPENDIX II. 



TABLES OF GASEOUS IMPURITIES, THEIR CHARACTERISTICS AND 

ELIMINATIONS. 

This enquiry is followed out in these tables in conjunction with the search for, or identification 
of, new lines in various of the gases : the main principle assumed being, that any new or faint line of 
any gas, ought to be more or lese visible in every tube, according to the visibility therein of the 
brighter known lines of that gas. 

With Hydrogen therefore (whose four standard lines there is no question about as to place), — 
I have begun its tables, with four double columns giving both a numerical expression for the intensity 
of its appearance, and a graphical reminder of the shape thereof, in each of the said four lines, in 
every tube observed. The Hydrogen tube, where alone these lines have full right to appear, has its 
number printed in the heaviest type. Other tubes which have hydrogen in their chemical composi- 
tion associated with something else, have their numbers printed in less heavy type : but the tubes 
where the alleged contents have no chemical claim whatever to hydrogen, and yet show its spectral 
lines, have their numbers given in thin type. 

Hence it may easily be seen, but with some surprise, that Oxygen and Ozone tubes show, besides 
their own lines, those of Hydrogen with even maxmimum force, though no Hydrogen ought to be 
there. Much Hydrogen appears also in Alcohol and Ammonia, but their compound formations claim 
Hydrogen as one of their constituents. Olefiant-gas and Marsh-gas have the same right to Hydrogen, 
but do not show so much of it in pure Hydrogen lines ; partly perhaps because some of it is retained 
with them to show the perfectly different lines of Carbo-hydrogen compound, or what is seen in the 
base of ordinary coal-gas flames. 

Turning then to, say the hitherto unclaimed line at 43 698 W.N. Place, and finding it strong in 
Hydrogen, Oxygen, Ozone, Olefiant-gas, Nitrogen, Marsh-gas, Alcohol, Ammonia, and all the tubes 
which have the known Hydrogen lines strong, — but absent in Chlorine, Cyanogen and others, when 
the known Hydrogen lines are either completely, or nearly absent, — we may say that we find the 
above hitherto unclaimed line appearing everywhere as a function of Hydrogen ; whence we seem 
entitled to claim it here as one of our new, low-temperature, lines of Hydrogen. 



GASEOUS SPECTRA IN VACUUM TUBES. 



143 





+3 

2 




O 




r=! 




-ta 




•iH 




£ 




n-J 




a 




<s 


C5 






co 




CD 

=5 


o 


Ph 


o 


O 

CO 


w 


PS 


Q 


o 
>> 


>H 


CI 




o 


w 


-^ 




C 




-Q 


W 


O 


w 


00 
CO 


Eh 


rH 


Fh 


1^ 


<J 


1— 1 
Ph 




«j 


<y; 




W 


a 
o 


fc 


T3 . 
03 -u 


(— i 


.3 3 


h^ 


sj to 

a <D 




c3 n 




X Q3 


£ 


03 r-±J 


W 


►3 o 


O 


.2 fl 


o 


03 >1 




co S3 


tf 


to 


p 




1* 


w" 



52; 
o 



sh 

03 

> 



O 



C 

03 
O 

w 

03 



03 

-JO 



o 
Ph 





■* 


OS 




r _ l 


CM 


,_, 


,_, 












^ 




.— i 






00 


to 










rH 


<M 






CO 


Ci 


■* 
















m 






CO 


CN 










^ 


"H 




rH 


rH 


o 


^ 












1—1 




rH 






i-H 


»H 










o 


O 




O 


o 


o 


o 












o 




O 






o 


o 










Ml 


-^ 




"* 


■^ 


^^ 


■* 












■^ 




■* 






Ml 


Ml 










- 


- 








■^ 














. 




— 






— 


■s- 
























































CO 


CO 




<M 


<M 




o 












CO 




<M 






(M 


i— l 










w o 


1 






00 














o 






















00 ! 00 








m 














00 






















00 ! 00 


<--. 






00 


e*> 












00 


















1 




en j o> 








Ci 














C5 






















CO CO 








CO 














CO 






















- 


- 








■*! 














— 






















w 


lO 








lO 














lO 






















1-1 


rH 








o 














rH 






















CO 


'CO 












t>~ 










to 










00 












to 


to 












i-~ 










to 










t^ 












CO 


to 


=«. 










5D 










to 










to 












C2 


o> 












C3 










OS 










C5 












CO 


CO 


1 










CO 










CO 










CO 












— 


— 


— 




m 


tfl 












lO 










lO 










lO 












^-1 


r-i 












o 










1-1 










"- 1 












to 


to 








o 














to 










CM 












to 


in 








(M 














to 










XO 












Ml 


mi 








■* 














-* 










-* 








-01 




o 


o> 








CJ 














C3 










c» 












CO CO 








CO 














CO 










CO 








CO 






„. 








S3 














— 






















1« 


m 






















m 










m 








CO 




1-1 


i— i 








rH 














rt 










1-t 








I-H 








A 




























































































<! 










































8 <N 

« CO 


jj 




















lO 






















3 


S5rH 


3 
0) 




















r-l 












M 










p. 

s 


|£« 


3 






















































































.2 








^ 




"™ 
















■■ 




^ 




^ 






. 




« 

3 

S3 

3 — 




■< 


™ 










l ^™ 






^^ 








■" 








™ 








So 
Js'oo 


lj 






t-. 


l>. 




m 




o 


o 


m 


to 








«5 








iO 




3 


w 


<* 


CM 


rH 




CN 




1—1 


a 


CO 


o 


« 


-<ji 


<M 


rH 


CO 


-* 




T-i 




Is 




ft 




■ 




— 




— 


— 





■ 


■ 


- 


^ 


■ 


_ 


^ 


■ 


H 


- 


__ 


















































oj 












































x: 


o irj 


>> 










































djCN 


3 4* 


CO 


o 


CO 


-* 


rH 
O 




<M 


CO 


CO 


>-4 


p 

CO 


^ 


o 


■* 


*r 


OS 


o 


ip 


lO 






ISCN 






rH 
















rH 






















5 


*» 












































c3 




3. 


- 


- 


- 


- 


— 


— 


- 


— 


- 


■ 


- 


" 


- 


» 


— 


■ 


■ 


_ 


— 




< 


o> 

iS o 


>> 














CO 


»o 




























Ph 1 ^ 


§ *ra 


00 


CO 


o 


«D 


CO 


Ml 


o 


CO 


tf) 


o 

rH 


!-H 


o 


CO 


>o 


CO 


O 


o 


•* 


CO 






is w 












































— i A 

E a 

a> bo 


6 

+ 
o 

+ 


o 




O 


O 

o 


5 


t5 


|2J 

o 


5z; 

o 


8 

K 


H 


HH 




rH 


O 




O 


CO 

o 


C3 
r?5 

+ 

o 


o 




P 


£ 




































W 
























r3 


■a" 






































^ 




o 


























3.5 
&r3 








"3 


'3 




03 
















-a 




















^ 


< 


o 




^^-^ 


N — 


s ~' 




- 








M 


CO 






M 




1 


c =3 


s-T 

<! 


o 

o 


'3 

o 

a 
a 


'3 

o 

■a 

o 


'3 

o 

■a 

6 


o 

o 


a 

Q 

SO 
o 

es 
>-. 

O 


d 

so 
o 

a 

C3 

o 


60 
O 

3 
>^ 
o 


o 

H 

-a 


a 

09 

SO 
p 

-3 
M 


oT 
o 


O 
in 

3 


09 
00 

p 

rH 


o 

rH 


o 

« 
09 

O 


09 
CO 

O 


oT 

a 
o 

N 

O 


0) 
r^ 
02 


09 

"8 



144 



PROFESSOR PIAZZI SMYTH ON 



•^3 
S 









o 
ft 



O 

o 

< 



(M 


to 




a 


CO 




-f 




_, 






o 


CM 


m 






CO 


M< 


CM 


r^ 






CO 






CO 


i-* 


00 




to 






(N 




to 






-* 


«* 


»n 


-* 




00 


OO 


& 5 


oo 


00 


(** 


«-~ 




1~ 






OS 


CO 


CO 






t^ 


00 


w 


i--. 








^o « 
































i-H 






T 


■* 


"* 


«* 


*<* 


■>* 




^* 






~f 


^ 


-* 






■* 


■* 


"* 


-^ 


















































o •":: 










































r. 










































- M 


















in 




\a 






ta 


\n 








03 


CO 


O 












































m 


CM 


-& 


CO 




m 






CM 


CO 


CM 






CO 


rH 


(M 


in 




-* o 


■* l^ 






to OS 


t— 00 


03 oo 












-* o 








CO o 










us .— 


to Ol 






CO 1^ 




Ol 00 












lO r-> 








t^ to 










CO in 


CO in 






CO -* 


CO ■* 


co m 












CO lO 








CO ITS 




















,_, 


,_ 












,_, 








,_, 










■* 


■* 






-* 


-r 


-* 












^3< 








«* 










= 


= 






= 


= 


= 












= 








= 










m 


in 








us 


o 












ia 








VO 










i— t 


1-1 






CM 


i— t 


I— ! 












1-1 








CM 












■* 


OJ 




OO 


to 










to 




o 




o 






o 










CM 


•Tf 




CM 


CO 










•* 




Ol 




CO 


















>H 




1-1 


o 










i-i 




l-H 




1—1 






r ^ 








,_, 


,_, 


,_< 




,_, 


,-H 










IH 




1— 1 




t— ( 






r- . 








■«# 


-^ 


^ 




■<* 


-* 










■<* 




■* 




Th 






-* 








— 


— 


— 




Br. 


— 










:Jii 








Hrf 














\d 


m 


o 




in 
















1C 










CM 








7— 1 


- 


-- 1 




iH 


rH 










Ol 




rH 




T-( 






O 








CO 


to 


t^ 




CO 


oo 














CO 


■* 


1^ 




CO 


"* 












to 




-* 


-* 














to 


CO 


OS 




o 


vn 








Ol 


o 


OS 




OS 


OS 














CS 


OS 


OS 




o 


OS 








i ° 


o 


o 




o 


o 














o 


o 


o 




!-H 


o 










-* 


-* 




■* 


-* 














■* 


^* 


>* 




■* 


•* 








j 


- 






- 


— 














- 


- 


Br. 




- 


— 








m 


lO 


o 




















KO 


US 






SO 


m 








C<I 


(N 


|— < 




CO 


CM 














CM 


CM 


I— I 




CM 


rt 
































tO ^ CM 










































to -^ CM 










































to 1--. OO 




























































o 


O 








































-* 


■* 






















«* 


















- 


- 






















" 


















o 










































CO 


CO 






















CO 


















OS 


CO 






t— 
















OS 


















CO 


CO 






CO 
















00 


















-<x 


^ 






-* 
















-# 


















o 


o 






o 
















o 


















- 


- 






SH- 
















Br. 


















in 


m 






















m 


















(N 


CM 






IN 
















CM 


















to 


00 






OS 
















to 


















oo 


OS 






o 
















CO 


















<M 


<N 


!-. 




CO 
















<M 


















O 


CM 






o 
















O 


















"* 


■«# 






>* 
















rh 


















_ 


— 






— 
















— 














T-H 






CM 
















- 


















Chemical 


3 




+ 

O 

+ 


o 
cf 


X 


o" 

O 


o 


o 




to 
o 


o 




X M 


X 

o 


» 


c 


X, 


O 


o 


to 

+ 
o 


o 






























































































































































































o 
































g 


rj 




OS 


■S" 


03 


<J 








o 






















•s. 




a 


o 














r/: 










■ O £ 




. 




ef 


^ 


o 




a 


^ 


a 


o 


a 


V, 

tS 


^ 


^ 


O 






OS 

OT 




Is 

55 


t 


<h 


o 


a 
o 

s 
s 
< 


'3 

o 

1 


a 
o 

3 
o 


o> 

a 
o 

O 


so 
o 
a 

eg 
>l 
O 


OS 

to 
o 
a 
ce 
>> 
O 


O) 

o 
fl 


o 

o 


Hydroge 
Iodine, 




a 

OS 

bo 
O 
H 


M 

O 
H 


4-» 

a 

ta 

O) 

O 


OS 

bo 
>■> 

O 


oT 

a 
o 

N 

o 


o 
rt 

^ 






GASEOUS SPECTRA IN VACUUM TUBES. 



145 









02 



o 
P3 

Q 
!* 

o 

w 

o 

< 

02 



■* 


CO 


i— t 


-* 


o 


CO 


in 








CO 


-* 




CO 


CO 




CO 


CM 


Ol 




^ 


OS 


or 


lO 


to 


o 




UO 








CO 


Ol 




05 


in 




CO 




Ol 




CO 


CM 


CO 


CO 


CO 


CO 


CO 


CO 








CO 


CM 




Ol 


CO 




CO 


CO 


CO 




CO 


^* 


^n 


-r* 


-* 


•* 


■<* 


-^ 








-fll 


-* 




■>* 


■* 




■* 


-* 


-*l 




•* 


"<* 


•* 


<«* 


-* 


-* 


-* 


-<J( 








■* 


■* 




->n 


■* 




^H 


-H^ 


>* 




kK 


— 


- 


— 


££ 


Si! 


22 


53 








^ 






- 


- 




- 








ZZ 


m 


m 


i—* 


o 


o 














o 




m 








m 


m 




o 


CO 


CO 


CO 


CO 


CO 


I-H 


•- 1 








Ol 


CO 




Ol 


CO 




CO 


•- 1 


'-' 




■- 1 


CM 


m 


Ol 


CD 


•** 


m 










CO 


Ol 




■* 


















CM 


in 


-* 




Ol 










CO 






Gl 
















•— ' 


iH 


»H 


i-H 


1— 1 


I-H 










I-H 


I-H 




O 






^,_ 










■* 


•»* 


■<* 


-sH 


-* 


■* 










-* 


rtl 




-Hi 
















■<* 


•>* 


-«< 


-^ 


•* 


-* 










^H 


TH 




-* 
















— 


" 


■ZH 


— 


- 


- 










- 


— 




= 
















m 


m 


m 




lO 












in 


in 




CO 
















Ol 


Ol 


'-' 


Ol 


■- 1 


•- 1 










r~t 


Ol 




■- 1 
















oo 


CO 


o 


1^ 


o 


CO 


CO 








CM 


CO 


o 


CD 


o 




CO 


CO 


t^ 






01 


Ol 


o 


CO 


Oa 


1^ 


o 








o 


Ol 


CO 


1~- 


CO 




-* 


oo 


Ol 






CO 


CD 


1~. 


l^ 


CO 


CO 


1-- 








1--. 


CO 


CO 


CO 


CO 




t^. 


CD 


CO 






CO 


CO 


CO 


CO 


CO 


CO 


CO 








CO 


CO 


CO 


CO 


CO 




co 


CO 


CO 






-* 


■* 


■* 


-* 


■* 


-# 


■>!< 








■* 


^H 


HJ1 


-* 


-31 




■* 


■* 


^* 






- 


- 


— 


- 


- 


— 


— 








- 


- 


- 


- 


ii* 




- 


— 


- 










in 




lO 


m 














CO 








in 
























































Ol 


-«l 


CO 


CM 


CM 








CO 


■* 


o 


CO 


CO 




CO 


CM 


CO 






»o 


■** 


I-H 


Ol 


Ol 


CO 


-* 










in 




M< 






in 




CO 






*# 


"** 


CD 


CO 


CO 


CM 


CO 










-**i 




O 






*^ 




CO 






US 


iCt 


m 


m 


m 


iO 


in 










m 




m 






m 




in 






CO 


CO 


CO 


CO 


CO 


CO 


CO 










CO 




CO 






CO 




CO 






"«^ 


•*a 


-* 


"* 


■* 


^ 


■"* 










-* 




-^H 






-^i 




■* 






— 


— 


— 


- 


- 


- 


- 










— 




h^: 






- 




- 






<M 


Ol 






CO 


Ol 


o 




















m 




o 






Ol 


■- 1 


1— 1 


I-H 


■- 1 










CM 




CM 






f-H 




^ 






, 


t-- 


I—" 


o 


CO 


oo 


i-H 








CO 


,_, 




t^ 






t^. 


Ol 


o 






lO 


■* 


-* 


t^ 


m 


CO 


-* 










in 




in 






t^ 


->n 


CO 






l-H 


iH 




fH 


i-H 


I-H 


i-H 








I-H 


i -1 




I -1 






I-H 


I-H 


I-H 






CO 


CO 


CO 


CO 


CO 


CO 


CO 








CO 


CO 




CO 






CO 


CO 


CO 






TJ1 


-* 


<« 


^H 


■* 


-di 


■* 








-*n 


m> 




-H^ 






■* 


-* 


'JI 






- 


- 


=*== 


- 


- 


- 


- 








— 


- 




— 






— 


- 


m 






CO 


CO 








m 


CO 








o 


o 




m 








-* 


CM 






CO 


CO 


CO 


I-H 


'- , 








Ol 


CO 




Ol 






CM 


* 


Ol 






o 


CO 


m 




Ol 












-* 


o 




in 






Ol 


CI 


m 














OS 












00 












Ol 


o 


CO 






oo 


CO 


CO 




t- 


<-.■ 


... 








1--. 


CO 




CO 






t^ 


CO 


oo 






Ol 


CM 


CM 




CM 












Ol 


CM 




CM 






CM 


Ol 


Ol 






^ 


~# 


■** 




-* 












-* 


■* 




-* 






-* 


-* 


■* 






= 




— 




= 












= 


= 










= 


= 


















CO 


Ol 






















o 


o 






(M 


Ol 


Ol 




Ol 












Ol 


Ol 




Ol 






CM 




■"• 






o o 1 

CO r-H 
O Ol | 


m 1 


, 




-* 












1~. 


in 




!>. 








o 








in 


CO 
















CO 


CO 




in 








1^ 




























I-H 






























































Ol 


Ol 


Ol 




Ol 












Ol 


Ol 




Ol 








Ol 








^* 


-* 




-* 












-* 


■* 




-<* 








^ 








- 


— 


■HH 




— 












-il 


- 




1 








= 








m 


^ 


m 






m 










m 


in 












in 








■* 


CM 




CO 


I-H 










Ol 


^ 




CO 








- 1 












o 


































03 




"cz C 




■3 


n 


























HI 
ft 






fc 




11 

6 I 


3 


+ 
O 
+ 


«3 

IB 


« 


o 

o 


O pn |z; 
o ° o 




(Z5 
o 


o 

IB 


w 


1—1 


a 


K 




O 


O 


+ 

o 

ft" 


O 
ft" 




















- 










































13 












































"3 














































































cr 


^^ 


fc> 


«1 
































3 


r2 


cu 


■o 




o 










<D 


m 
















eS~ 


< 


o 


a 




CI 


o 


a 




d 


^ 


o 













o is 

la 






" 
"o 


PI 

o 


'3 

o 


'3 .£ 

o t. 


cu 


CD 

bD 
O 

a 


bo 
o 
P! 
as 


o 

o 

Eh 


bo 
o 

Eh 

n3 


CI 


r-t 


a> 
o 


p 
o 


"3 

c« 
cci 


a 

0J 

bD 


aT 
d 








3 


o 

^ 





t4 

o 


o c 




o 


IB 


IB 


o 


fc 


k 


01 

O 


O 


o 


73 





VOL. XXX. PART I. 



146 



PROFESSOR PIAZZI SMYTH ON 









m 
(— i 

w 

o 
o 
ft 
A 

K 
o 

w 
< 

J/} 





_ 





















































































































OJ o 


r-* 


bO 








■<*l 


CM 


CN 








o 


o 




o 










\a <M 


CO 


o 








CO 


o 


CO 








<N 


o 




U-w 










to OJ 


oo 


+2 








CO 


cs> 


CO 








o> 


en 




00 










<-» t-- j t^ 


fc 








f^ 


i^. 


1^ 








1^ 


i^ 




l^ 










-^ "* "*»* 








-* 


^> 


-* 








-* 


■* 




■* 












« 






































o 




































* 

O 


— 










— 


— 


— 








— 


— 




— 










in 


o 










o 












»o 


o 














PH 


r— < 










"- 1 


rt 


C-l 








^ 


^ 




CN 










CO 


CO 


CO 




CO 






CO 








CO 


CO 


i^. 














-** 


-^ 


J~ 




in 






M- 










■"*< 


CO 














>-i 


■"" 


rH 




■"■" 






rt 








^ 


•"* 


r ^ 










+- 




i— 




t^ 




t-- 






I— 








1^ 


(N« 


1^. 














-* 


^f 


-* 




■* 






^* 








-* 


^ 


-# 














— 


— 


-H- 




— 






" 








— 


— 


— 














>o 


VO 


o 




o 






o 








lO 


>o 


o 














r-l 


rt 


I— < 




1— 1 






"* 








r-H 


rt 


<M 














CO o 


CO 


CO 




^h 






en 






Oi 


■* 


o 


CO 


i>~ 


O) 










*-* CO 


t^ 


1- 




CO 






CO 






i^ 


VO 


00 


1^ 


CO 


OO 










l~- CO 


CO 


CO 




CO 






00 






CO 


OO 


oo 


00 


CO 


CO 






* 




CO co 


CO 


CO 




CO 






CO 






CO 


CO 


CO 


CO 


CO 


CO 










"«* "^ 


>* 


■•3< 




t* 






TK 






-* 


■* 


•* 


^* 


-* 


■* 










-*-> 








































cS 








































c — 


— 


-*. 




-~ 






- 






— 


— 


— 


— 


— 


-H- 










o >o 


o 


<M 




>o 






o 






c 


o 


lO 


o 














R,H 


CM 


r-H 




■- 1 






CO 






'-' 


C-l 


rt 


CM 


-* 


(N 










CO 


05 




















o 


CO 


CM 






.—1 


OS 


CM 




c» 


c 




















CO 


OS 


-** 






o 


CO 


T— 1 




CO 


o-. 




















OS 


CO 


00 






Oi 


Oi 


CTi 




in 


J^ 




















>o 


tfi 


\a 






lO 


>o 


IO 




-» 


«* 




















^*< 


■* 


■* 






^ 


■* 


^f 












































" 








































>o 


lO 






















in 


\a 














■- 1 


T-H 




















(M 


■- 1 


*-< 






(M 


CM 


"- 1 




. ! o 


»o 




CO 














CO 


CO 


CO 














oa o> 


oc 




cc 














OS 


CO 


CO 














«5 CO 


^D 




co 














CO 


CO 


CO 














-* | -+■ 


-* 




-r 














-tH 


** 


-* 














"^ "^t" 


■»* 




T 
















"* 


-f 














— 


— 


— 




— 














— 


— 


— 






















ica 














lO 


















CM 


CI 


f-H 




















■M 


<M 














1^ 








CO 








o 








1^ 


















-* 






f 








C-l 








•* 
















VO 


l£3 






lO 








>o 








lO 
















-* 


-fl 






-r 








■* 








M< 
















-* 


■^i 






^f 








■^ 








-* 
















— 


— 






— 








— 








— 
























in 








c 
























CM 


CM 






















Ol 


















ci 
































C3 




■3.2 


a 


r> 






























!z; 




6 I 


+ 
o 
+ 




HI 
SZ3 




o 


C 


Szi 


^ 


o 


o 


h- 1 


i— ( |u 

o 


SzS 






+ 
o 


o 








































0. . 


















2 


r^f 


















- 5 








. 


rf 




^~ 


; 3* 


o 

>> 


< 


















o jz 








O 


X 




c 


o 


s 










"C 


cr" 








3 — 








*-" 


O 










o 




oT 




M 


a 




t- 








r^ 








« 


13 


a 


a 


rt 


" 


f 1 


O 




w 






< 


"o 

3 


c 
o 

E 

■5 


o 

o 


'3 

o 

r? 
O 


o 


eg 

60 

o 
13 
CS 
>^ 

O 


o 

d 

o 


bo 
o 
a 


p 


OJ 

ho 
o 


.2 s« 
"s 3 

O ■— 1 
i— ( r^\ 


o 


CO 

p 
o 

-4-> 


1 C •" 

* IS S 

o o o 


-t-J 













































GASEOUS SPECTRA IN VACUUM TUBES. 



14/ 









O 

o 
o 

W 

o 

«2 



00 


oo 




















00 


to 






1—1 










Ifl 


m 




















lO 


CM 






as 










CO 


CO 




















eo 


CO 






CO 










in 


m 




















m 


lO 






m 










Ui 


in 




















m 


IQ 






m 










w 


W 





















m 


lO 






m 











■- 1 


rH 




















'-' 


"- 1 






- 1 










as 


OS 


I-*. 


















as 


t--. 






o 


o 








a. 


OS 


ex 


















OS 


OO 






OS 


CN 








i^ 


*- 


x^- 


















*>• 


t-- 






1^ 


00 








--41 


-* 


-* 


















-t»i 


-* 






-<f1 


-rfl 








in 


m 


iO 


















o 


m 






m 


in 








— 


— 




















— 


:=h; 






— 










O 


o 




















o 


o 






o 


^ 








<n 


CM 


r-i 


















CI 


I— 1 






<N 


o 
















^ 
















^ 


^ ^ 

















CO f~- 


o to 


to t>- 


CM -Hi 




in 












eo t^ 


co as 






1^. to 


o 








to CO 


w -* 


(N <M 


1— CO 




o 












to CO 


to o 






in eo 


OS 








-hi in 


-Hi in 


« O 


-* in 




m 












if m 


-31 in 






-ji in 


-* 








1— 1 rH 


I— 1 t-i 


*-H r-« 


1-1 1— 1 




,_! 












t-H l~t 


y-i •— < 






i— i i—i 


1— 1 








m m 


m m 


to m 


in in 




m 












o m 


>n in 






in m 


in 








m 


— 


-^ ^. 


= 




=Hr 












m 


1 






= 


— 












o 






© 































cm 


CM 




- 1 




- 












« 


<M 






CM CM 


• H 








OS 


OS 




00 
















as 


CM 






OO 
















to 
















-o 


00 






CO 










to 


to 




to 
















to 


to 






to 










o 


o 




o 
















o 


o 






o 










in 


ire 




m 
















m 


m 






m 










- 


— 




















I 


\ZS 














- 


in 


lO 




o 
















u-5 


CO 
















CM 


cm 




CM 
















Ol 


1— 1 






CM 










CO 


m 


H 












to 


OS 




o 




















OS 


60 




































as 


Oi 


a 












OS 


o 




OS 
































as 


o 




cn 


















-h< 


-HI 


o 












■* 


m 




•-* 






















ft* 

o 




































^ — 
















^H 




■■« 




















~~^ 






































CM 


cm 














<M 


o. 




•M 




















c~ 










































_. as 

S to 

x o 

g°3 












O 


>o 


eo 


o 










un 


-* 


CM 


















U7) 


CO 


OS 


OO 










to 


to 


rH 




CO 


CO 












CO 


CO 


CO 


CO 










CO 


CO 


-* 


















OS 


OS 


as 


OS 










OS 


as 


OS 




■* 


-HI 














-11 


" 
















-* 


-«1 
















































o o 












__ 


































































lO 


% 














u^ 


lO 














m 






1-1 


I-l 














CO 


(M 


o 


T— I 










1—1 


o 


^ 






■ 


































C3 




z i 


1 

c 


+ tn 
O -4, 




o" 

O 


O 
O 


o 


o 


O 


o 


O 


S "" ' 


o 


fc 


o 


O 


6 


O 


+ 


O^ 


C ' 


1 


+ o 


































c 


1 


fe 












































































s 


















^H 


T3 


































.. 




o 






















1 
1 

■ 

1 
1 
1 






3 


as 




5 


"o 


>1 


<1 

o 








3 


tn 










i; 

'*3 


O 


3 

S 

g 
<3 


_o 

'3 

o 

J? 
3 
o 


o 

°3 

o 

.p 

s 

a 


aT 

a 

o 

o 


a> 
o° 

CI 
O 


o 

a 
a 


to 

o 

a 

C3 
t*s 

o 


o 
p 


1J 
60 

p Zl 


zn 

o 


R 

as 
bX) 
p 


o 

us 

3 
p 


O 
+^> 

s 

td 

as 
O 


as 

>s 

X 

O 


aT 

a 

o 

NS 

o 


as 

«2 













































148 



PROFESSOR P1AZZI SMYTH ON 






« 

X! 

O 

o 

w 

o 







-*.. 




















































































co 


CO 






*»< 




























CO 








CM 


<M 


































CM 








O 


O 


O 




•^ 




























O 








CM 


CM 


— 




CM 




























CM 








■>* 


** 


■9 




«« 




























■* 








ss 


— 






- 




























— 








lO 


m 


































in 








*"• 


"- 1 






eo 




























rH 






























e~. 






















o 


CD 


















Oi 


CM 


■** 












O 


m 






ce> 


CO 


















** 


m 


i-t 












CO 


o 






CO 


CO 


















co 


CO 


CO 












CO 


CO 






Oi 


o> 


















o» 


OS 


Oi 












oa 


Ol 






CO 


CO 


















CO 


co 


CO 












CO 


eo 






— 


















i 




— 












- 


— 


























CO 
























CO 


CO 


















rH 


l-H 


CM 












CO 


CM 






o 


















o 






o 












o 


^*< 






t^ 


CO 
















in 






eo 












t^. 


© 






CM 


CM 
















CM 






CM 












CM 


CM 






to 


CO 
















CO 






CO 












•o 


CO 






CO 


CO 
















CO 






CO 












CO 


CO 






— 


— 
















— 


















— 


— 






in 


m 






















CO 












m 


o 






,H 


H 
















- 1 






o 












rH 


i-H 






o 


o 






















CO 












o 








o 


Oa 






















rH 












o 








co 


m 






















CO 












CO 








CM 


CN 






















CN 












CM 








CO 


CO 






















CO 












CO 








- 


- 






















- 












- 








m 


m 






















eo 












m 








o 


o 






















© 












o 










eo 












































oto 

S m 

C- rH 


.f 
















m 








»>. 










U5 






_ 


fc co 


1 
















o 








o 








CM 


CN 






o 


gw 


^ 








































V 
























































































S3 














































co 












































g Oi 


>-, 








































o 


5m 


s o 




o 


o 


o 


o 








in 




o 


in 




o 


o 






e-. 




£ s 


55 «^- 


g 0* 




rH 


»H 


rH 


r-l 








o 


CM 


H 


rH 




CM 


rH 


^ 


'S' 


-> 




In 

;- a 


£"* 




















































































a to 












































g-o 


g OS 
5 CM 


>-> 
JjrH 




O 














m 










o 












c 




te: co 


















|H 


CM 






■* 


CM 




CO 


CO 






C) 














































>> 








































r* 


P-t(N 


3 O 




o 


m 


o 


o 


lO 






m 




o 




o 


o 














fc»H 




Gfl 


©J 


<M 


■H 


cn 






CM 


T* 


■H 




CO 


CM 




<tf 


m 








>^ 






















































































signsition. 


CJ 

•a 

+ 
o 
+ 


O 

CO 

o 




o 
o 


o 


3 


55 

O 




55 
O 




« 


l_l 


o 


fc 


o 




o 


o 


ca 
+ 

o 


O 


w „ 


525 




































a 










































































































,_ri 


■73 








































O 
































_r 


,§ 




f 


^ 


>> 


"Sj 










oT 




























































o 


X 






o 


fc» 






























«j 


o 




*■ — 


*— ^ 






„ 




»r 




X 


rt 






t-, 




Si « 

- V 


3 


.a 
o 
o 

3 


o 



3 

-5 


"8 

o 

4= 

H 

6 


'3 

o 

42 

3 
o 


qT 
a 
'£ 
o 

O 


to 
o 

a 

C3 
>> 

o 


o 
to 

o 

a 


CO 

bO 
O 

a 

« 
>> 
O 


O 


0J 

bC 

o 

-a 


a 
o 


C3 

H 
C3 

a 


a" 

CO 

60 
o 

>5 


CO 

a 
o 

55 


o 

a 

cS 

ca 

CD 

o 


a" 

CD 

V 

o 


a 
o 

NJ 

O 


r* 

-t-> 

"3 

CO 


•5 



GASEOUS SPECTEA IN VACUUM TUBES. 



149 



^3 

ss 

•<s> 

s 



to 



o 
o 

PR 

w 

o 

CO 



1 

o ; 


1 




h* 








CM 


o 










t-~ 








o 






to 


CM 






to 








Oi 


o 










to 








o> 






CM 


'- 1 








o 








O 


'- , 










o 








rH 








I— < 








rH 








rH 


rH 










rH 








rH 






,_, 


m 


-T- 






lO 








m 


m 










m 








>o 






m 


— 








— 








— 


— 










-S 








— 
























lO 












t^ 
















rH 








CM 








1-1 


CO 










■- 1 








rH 






rH 


o 




I— 1 










oo 














•* 


t^. 






O 








CM 




lO 










-* 








t^. 






CO 


^rji 


to 




CM 








W 




in 










in 








m 






« 


i£J 


>n 




m 








to 




to 










to 








to 






to 


to 


to 




to 








■<* 


* 


-* 










-* 








■* 






-* 


-* 


•>»< 




■* 








— 





-X 










- 








— 






-s 


-* 


-S 




— 








o 


ta 
























>rs 
















rH 




i— i 










lO 








rH 






rH 


-* 


<M 




•- 1 








00 


Oi 


































00 


CM 


oo 




O 


o 


































o 


OJ 


CM 




t -1 


•— * 


































rH 


o 


rH 




to 


to 


































to 


to 


to 




■* 


-"»! 


































-<* 


-* 


■^ 




-TT-. 


— 


































— 


— 






o 


lO 










































rH 


rH 


































"- 1 


CM 


•"' 




O 




































o 


<M 






H< 




































-* 


•^ 






to 


to 






















* 












to 


to 






<N 


CN 


































CM 


CM 






-3< 


■«* 


































-* 


"* 






-^ 1 _ 


































— 


— 






W 


r>. 


































!>. 


lO 






"- 1 


- 1 


































rH 


rt 








o ?o 


- 










































.5 »o 


R 
















U3 

o 








b. 








cq 


eg 






-tf 


£ CO 










































o 


(S* 


•■= 
















































































































































































>> 








































. g IB 


5 *n 
Ph to 


2 ? 




O 


p 

rH 


p 

rH 


p 








o 


im 


p 


>rs 


=.. 


p 

CM 


p 


^ 


^ 


r-< 






E5 t". 






















































































o -— 


fe 


M 
































































































































«e to 












































£.*»* 












































p.iS 


S <M 


'3 <S» 




o 














m 










O 














fc to 


g rH 




rH 














rH 


(N 






-* 


et 




CO 


CO 








^ 










































5 


(0 












































S"* 


t". 










































(h <M 


■3 © 




O 


"3 


O 


o 


W 






VO 




o 




o 


o 














fc<-l 




CN 


C» 


CO 


rH 


CM 






CN 


•>»< 


rH 




CO 


CCl 




^ 


m 








kJ -* 


a 










































& 


" 










































o 

•a 
+ 
o 
+ 


o 

CO 


rZi 


o 


O 

O 


o 






r3 

O 




w 


- 


o 


& 


O 

r25 




o 


o 


+ 
o 

W 


O 












































o 




















r 3 






















^S 








-rj 


2 




0> 


P? 


>> 

S 


<i 










0J 












•^ 3 
















o 
















W 


















<1 


o 




' 






o 
p 

>> 

W 










M 


c3 






Fh 




Names 
seurc 




o 

H 

o 


o 

g 

S 


cj 

o 

.a 

rH 
cS 
O 


o 

8 
o 


3 
o 


CD 

bo 
o 

3 
&-> 

a 


a 

aj 
bO 
o 

3 
>> 
Q 


s 

bD 
o 
a 

>> 

O 


bC 
p 

-5 
>> 

W 


o 


o 

CO 

CS 


05 

o 


O 


o 

c3 

trt 

a> 

o 


R 

CD 

60 

o 


oT 
d 
o 

o 


+3 

cyj 

















































c» 

> 

C 
c3 





•°.t» 




-h Sh 




CD -rH 


fl 




o 




ci 


r,, O 


r*S 


03 ." 


o 


g >» 


o 


CO Q 


.3 


2 § 


o 


+3 ° 


S 
S 


*^ o 


rt » 


<J 


CB O 


in 


Hd'" 


en 


S2 -73 


o 


2 ^ 


in 


CO M 




rS P. 


— 


tc P^ 


— 


0) S3 






CN 






r-H O) 


4- 






a) o 




— «*H 




.S cu 




i~* rH 




o o 












1 1 -f-< 




O aj 




-3 ^ 








CCS QJ 




« CO 




CO cu 



rH 


m 


o 


cu 






EC 


.-' 


UJ 


ccj 


fl 


+^ 












<-. 


fl 


s 


o 


4~> 


rO 


O 




cti 




CU 


ro 


H 


•n 












C3 


ro 


.o 


4-> 






r^ 


P 


O 


CO 




-j 


f/) 


rH 


flJ 




fl 




1 1 


a 


a 


^> 


CD 

bo 

o 




tH 




-M 




• rH 




^ 




I* 




QJ 




a 




?-t 




O 




<4-) 




r^! 




o 




rH 




r« 




CD 




Oi 





150 I'ROFESSOR A. S. HERSCHEL ON 



APPENDIX III. 



ON THE MAGNIFICENT FEATURES EXHIBITED BY END-ON VIEWS OF 
GAS-SPECTRA UNDER HIGH DISPEKSION. 

By Professor A. S. Herschel, M.A., F.R.A.S. 

I have enjoyed rare and exceptional opportunities during the last two years of often 
heholding, although with very little leisure for studying thern minutely, the splendid spectacles 
presented by fluted and other spectra of incandescent gases in the optically exquisite, and 
surpassingly powerful compound-prism spectroscope erected by Professor Piazzi Smyth for 
examining Aurora?. 

My first views through the instrument, in April 1878, were chiefly confined to detecting 
and observing impurities of the rarefied gases said to be contained in the fine series of lateral 
view Geissler-tubes which Professor Piazzi Smyth had obtained from the late Mr Geissleu 
himself, for his investigations. The admirable disposition of the pointer, and means of pro- 
ducing and shutting off immediately, close to the upper and lower edges of a spectrum under 
examination the spectra of comparison tubes, combined with the extraordinary dispersion and 
transparency of the prisms, and the precision and solidity of all the movements and adjust- 
ments, made this first reconnaissance of the apparatus no laborious investigation, but on the 
contrary a brief enjoyment of the most unexampled luxury of ease and celerity in ocular 
discriminations which can very well be desired or arrived at with the spectroscope. 

Among my numerous notes of this first acquaintance, it will suffice to mention as an 
important observation, that the bands of carbon received especial attention ; aud that their 
individual variations in strength from tube to tube (for scarcely any tube seemed to be free 
from them), unaccompanied by any visible alterations of their positions, is a subject which well 
merits the foremost share of recognition in a circumstantial description. 

It is well known that the blow-pipe-flame green, citron, and orange bands differ in the 
spectral places of their leading edges and shaft-lines from the corresponding carbon band-edge 
positions observed with the extraordinary ubiquity just described in vacuum tubes. Super- 
posed upon the vacuum-tube carbon-bands they yet in general also exhibit their comple- 
mentary array with a varying degree of strength, more or less prominently, as intruders. These 
two distiuct orange-citron-green confederations certainly have an independent origin. One of 
the two is absent and not at all discernible in the blow-pipe-flame, while in almost every 
vacuum tube it can either be traced perceptibly, or it is even troublesomely conspicuous ; and 
it is especially resplendent in tubes of carbonic oxide and carbonic acid. If it is anywhere 
very much subdued, it is so principally in an olefiant-gas vacuum tube, where the tricoloured 



END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 151 

band-system of the blow-pipe-flame, on the contrary, supersedes it almost entirely. In other 
tubes the two systems are simply superposed upon each other, or mingled together in various 
proportions of intensity. 

Similar to the independent variations of the two orange-citron-green band-combinations, I 
noticed a marked character of individuality in the dark blue blow-pipe-flame band at the solar 
line G, which with its faint precursor, and with one strong blue band (near F) between it and 
those two sets just noticed, also constitute together very persistent features of carbon impurities 
in gas vacuum tubes. This most refrangible blow-pipe-flame band makes its appearance to- 
gether with a prodigious development, just following it between G and H, of the six or seven- 
rayed violet line-cluster which Professors Liveing and Dewar have recently ascribed, in a 
Paper presented to the Royal Society of London, to Cyanogen, with extraordinary luminosity 
in a marsh-gas tube. The latter line-cluster I have observed as a single and solitary lucidum 
in the beautifully blue arc of flame between the pure carbon poles burned in the Brush's, or 
Anglo-American Company's Electric light. Its freedom in each of these two cases from any 
simultaneous traces of the less refrangible band or cluster midway between H/3 and H7, also 
referred by the same writers to Cyanogen, but which I have never yet detected in gas vacuum 
tubes, makes me doubt the correctness of their interpretation that it belongs to cyanogen, 
and to venture to attribute the six-lined blue-violet clusture just beyond G, and perhaps 
also the most refrangible band at G of the blow-pipe-flame spectrum, to the incandescence of 
marsh-gas. 

Ably supported as the assumption is, no doubt, that there exist low-temperature spectra 
of the chemical elements, particularly of the metalloids, and sound as some of the evidence is, 
without question, by which the important theory has been established, yet the identification of 
the low-temperature spectrum of carbon, if it exists, cannot be said yet to be unanimously 
represented as accomplished. The independent radiancies of the several individual bands and 
hand-combinations which together constitute the carbon impurities of gas vacuum tubes, 
including that presented in the blow-pipe flame, are so strikingly various and unconnected, 
that a choice among the band-series produced respectively by olefiant-gas, by carbonic oxide, 
by marsh-gas, by cyanogen, and it may be by other carbon compounds, is one of some 
difficulty, before it can be positively affirmed which of all these is the low-temperature 
spectrum of elemental carbon by itself. That several spectra of different degrees of tempera- 
ture may exist, will scarcely explain the predominance under the same conditions, of three 
different spectral systems in such tubes as those of carbonic oxide, olefiant-gas, and marsh-gas, 
nor for the arbitrary admixtures of these three separate systems which high dispersion and 
accurate measurements easily detect as present in various abundances as common impurities in 
ordinary gas vacuum tubes. 

The following measurements of the linelets and of some intruding shaft-lines in the 
citron-band of a carbonic acid tube, made a year later with the improved tubes allowing end- 
on vision, but with the same prisms and micrometer-screw of the aurora-spectroscope, will 
show the precision of detail of which the instrument was capable, and at the same time the 
regular aud definite character of these two kinds of carbon lines and bands which present 
themselves in vacuum tubes as intruders one upon the other. A translation of the readings 
into wave-numbers per British inch, made at the time, although not possessing the accuracy 
which Professor Piazzi Smyth's later conversion tables for the instrument's readings would 
have given them, is added to the measures of the list. A certain regularity of the intervals 
among the linelets (although not among the intruding lines) is discernible, which may, perhaps, 



152 



PROFESSOR A. S. HERSCHEL ON 



not be only apparent, but may have a natural signification. This easy measurement with high 
dispersion will serve as a small pendant to the vast stock of observations with lower power set 
forth by Professor Piazzi Smyth in the foregoing paper, in illustration of the prodigious mul- 
titudes of details observable in a single shaded band, with the spectroscope's utmost resolving 
power. 

In comparison with other observations it may also, perhaps, suggest hypotheses of some 
slight use and interest for future explanations, in the confident hope which may now be fairly 
entertained, that the speeding advances of theory and observation will at no very distant time, 
by their joint discoveries, penetrate the physical meaning, and interpret the beautiful chromatic 
harmony of these close-ruled spectral bands. 



Micrometer Readings 
(revolutions). 



Intrusive Blow- 
pipe Citron Lines. 



CO., Citron 

Hand 

Linelets. 



Wave-numbers to an 

inch. 



28-500 (citron 
line, 1). 



29-360 (citron 
line, 2). 



28-893 
•950" 
■978 

29-008 
•035 
•068 
•100* 
•138 
•177 

•225 
•277 
•335 

•395 
•470 
•538 

•616 
•702 
•796 
•894 



Lines and 
Linelets. 



(45,066) 
45,260 
285 
300 
312 
325 
340 
356 
373 
390 



410 
431 

457 

(45,468) 
482 
514 
542 



576 
612 
651 
691 
733 



Intervals. 



25 

15 
12 
13 

15 
16 

17 
17 

20 

21 
26 

25 

32 
28 

34 

36 
39 
40 

42 



Average 
Intervals. 



10 intervals 
of 13-0 
each (?; 



6 inter- 
vals Of 25 "3 
each = 2 x 
12-67. 



5 inter- 
vals of 38-8 
each = 3 x 
12-93. 



Micrometer Headings 
(revolutions). 
[continued.] 



Intrusive Blowpipe C %^° n 
Citron Lines 



(continued.) 



30-245 (citron 
line, 3). 



31 -000 (citron 
line, 4). 



31-700 (citron 
line, 5). 



Linelets. 
(contd.) 



29-994 

30-098 
•205 

•325 
•448 
•575 
•700 
•850 



31-124 
•280 
•440 
•605 

•777 

•930? 
32-100 
3-2-323? 



Wave-numbers to an 

inch. 

(continued.) 



Lines and 

Linelets. 

(continued.) 



45,733 



45,778 
825 
(45,840) 
875 
924 
977 

46,026 
084 
136 



(46,146) 
196 
258 
323 
386 

(46,420) 
452 



Intervals. 
(contd.) 



508? 
571? 
656? 



Average 
Intervals. 
(continued.) 



45 

47 

50 

49 
53 

49 
58 
52 

60 

62 
65 
63 

66 

56? 

63? 
85? 



The linelets end here in haze. 



3 inter- 
vals of 5037 
each = 4 
12-59. 



5 inter- 
vals of 63 '2 
each = 5 x 
12-64. 



Citron-hand in an end-on C0 2 tuhe, with intrusive blowpipe citron lines. Prism 9 ; dispersion 33° from A to H. 
April 1879. N.B. — The linelets of this group from wave-number 45,400 onwards (width 1 or 2 wave-numbers), are 
really exceedingly close pairs, opening gradually in width to 30 wave-numbers apart at last, but each pair is only noted 
here by its mean place, as if it were a single linelet. 

A similar set of measurements to these was taken very rapidly in July 1879, of the fluted 
spectrum of Nitrogen, a translation into wave-numbers of the excellent Table of that spectrum 
in Angstrom and Thal^n's Memoir on "The Spectra of the Metalloids," f having presented un- 
mistakable indications of an arithmetical progression in the wave frecpaencies of its lines in the 
red to green portion of the spectrum. Tubes of sufficient purity to show this Nitrogen close- 
fluting or serration without interruptions or obliterations from the red to the green end, are 
however of rare occurrence ; and neither those of air, nor of Nitrogen and its oxides presenting 



* Dull and band-like ; probably double lines (?) 
t Nova Acta Reg. Soc. Sc. Upsal., Ser. iii. vol. ix. 



END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 153 

good end-on views of it, I had abandoned the project of repeating the Swedish experimenters' 
measurements, and I obtained instead some rapid measures of a superb series of red and yellow 
flu tings in a new end-on cyanogen tube, and had completed its astonishing survey before it 
occurred to me to compare the series with the imperfect but very similar colonnades visible in 
the other nitrogen-containing tubes. A rather weaker display seen in the tube marked 
" Nitrogen " was measured, and it was immediately seen to be identical with the series in the 
cyanogen tube. The identity of the same series in all the compound-of-Nitrogen tubes which 
I have since tested also leaves no doubt of the absolute constancy of its appearance, as far as 
it is visible without confusion and obliteration by other substances in these tubes. Thus the 
wished-for end of its remeasurement was already attained in the set of readings noted of the 
magnificent array of linelet groups seen ruling the red and yellow portions of the cyanogen 
spectrum.* 

In this marvellously beautiful array of triplets (the tube has now lost its original perfec- 
tion), one following another at a little interval, a small intrusive line in one of the intervals 
did not conform to the measures. A suspicion of its origin being thereupon excited, the hydro- 
gen comparison tube was lighted up, and it immediately proved to be Ha. A mere trace of 
hydrogen so weak and feeble as this was, is, from the ordinary prevalence of aqueous vapour, 
rather an exceptional occurrence in a vacuum tube. The conclusion, however, which may be 
drawn from its scarcity in this instance is a point of special moment to the theory of these 
channeled spectra, the explanation of which was given to me on the occasion of this occurrence 
by Professor Piazzi Smyth. As the proposal to contribute this Appendix has prompted and 
invited me to the fullest freedom of communication, I gladly avail myself of the liberty with 
which I am thus entrusted to reproduce it here, in order to show upon what small and 
apparently insignificant appearances, sound and just views of the nature and origin of gaseous 
spectra may sometimes come to be correctly founded. 

The absence or deficiency of hydrogen is demonstrative of the sensible absence or remark- 
able deficiency of aqueous vapour, and consequently (admitting the purity of the included gas from 
air) of disposable oxygen in the cyanogen tube. Yet not only is the serried nitrogen colonnade 
most resplendent in its electric spectrum, but so also are those vacuum tube carbon bands which 
are best known as constituting the electric spectrum of carbonic oxide and carbonic acid vacuum 
tubes. Of all the lateral and end-on tubes examined, these latter bands, like those of Nitrogen, 
were noticed by Professor Piazzi Smyth to be most brilliant in this Cyanogen one, in which yet 
there can hardly be assumed to exist more than a mere trace of Oxygen set free, to combine 
with the carbon and nitrogen, by the electric spark ! A new Cyanogen tube supplied to me quite 
recently by M. Salleron yields an electric spectrum of the very same description. The opinion 
held by Angstrom and Thalen, therefore in their Memoir appears to be scarcely tenable, that the 
fluted and banded spectra just mentioned are those of the oxides of Nitrogen and Carbon, but 
it seems more probable that these are in fact the true low temperature spectra of those metal- 
loids. The coal-gas or blow-pipe flame spectrum, on the other hand, is probably attributable to 
olefiant-gas.f A similar band spectrum is recognised by Professors Liveing and Dewar as 

* Salet, as well as Plucker and Hittorff, struck by the identity of this spectrum in all the nitrogen-bearing 
tubes, was led to the opinion from its constancy that the real source of the fluted spectrum is nitrogen itself. 

t A different view of this spectrum is, however, taken by Mr Lockyer, in whose opinion it is one form of the 
spectrum of elemental carbon. The smooth-shaded tube-carbon bands, in fact, resolve themselves into the line-bearing 
gas-flame ones on simply strengthening the induction discharge with a condenser, and especially on introducing at the 
same time an air-break also in its course. The experiment was tried after the present paper was read, with the above 
described carbonic acid and cyanogen vacuum-tubes, on July 23, 1880, on its prescription by Mr Lockyer to the writer 
of the paper, and to Professor Piazzi Smyth, and it succeeded in the presence of its suggester, literally as he expected ! 

VOL. XXX. PART I. Z 



154 PROFESSOR A. S. HERSCHEL ON 

belonging to Cyanogen (unless it may be due partially to marsh-gas), and the proper spectra of 
the oxides of Carbon and Nitrogen if these gases exist at such high temperatures undecomposed 
may in the course of further trials and examinations of the spectra of ignited gases eventually 
come to be discovered.* Multitudes of fixed bright lines in the spectrum of vacuum tubes 
enclosing pure hydrogen, are confidently regarded by Professor Piazzi Smyth as constituting 
together the low temperature spectrum of hydrogen ; and it is assumable that as no attempts 
to produce the spectrum of aqueous vapour in vacuum tubes have yet been attended with 
success, so also the oxides of the metalloids may be too easily decomposable by the electric 
spark to allow the natural spectra of the oxides of Carbon and Nitrogen to be easily exhibited 
in Geissler tubes. 

New end-on tubes of exceedingly hard glass, made and filled this year for Professor Piazzi 
Smyth by M. Salleron, have afforded new means of measuring the least refrangible section of 
the nitrogen channeled spectrum in Nitrogen, Nitric Oxide, and Cyanogen, with the advantage 
of the fuller conversion tables now constructed for translating into British wave-numbers the 
readings of the several prism-combinations of the aurora-spectroscope. 

The general appearance under high dispersion of each successive escarpment of the 
nitrogen serration is that of a double-notched band,f whose two teeth or bright edges face 
towards the red ; on the downward or fading slope (towards the violet) of each toothlet's descent 
are two lines besides the leading line at the edge, of diminishing brightness like the haze on 
which they lie. They divide the first slope into three apparently equal parts, while on the 
second slope the two lines are to appearance similarly placed, but the slope extends to once or 
twice their joint range further, before it fades out and leaves a dark space of a little breadth 
before the same double-notched escarpment begins again. There are also two lines preceding 
the new escarpment edge very similarly spaced asunder, and from the edge, to the other linelets 
of the group ; and this space between two adjacent linelets is only a fifth part wider (in wave 
frequency) than the interval between the two chief sodium-lines (Naa 1( a. 2 ). It embraced on the 
average six divisions of the micrometer screw-head, in the No. 9 prism, whose repeated read- 
ings of the same line seldom varied so much as one division, and in general only a few tenths 
of a division of the screw; but haziness, expansion, division, and supplanting of the lines due 
to coexisting impurities of other spectra in the tubes rendered exact readings of the fainter of 
the above lines mostly difficult and sometimes impossible. The only good measurements 
secured and entered in the accompanying Table were those of the three linelets on the first tooth, 
together with the leading Hue only of the second tooth. The results for the remaining lines 
lying between the second tooth and the first tooth or edge of the following escarpment, are so 
devoid of regularity, that, perhaps from their faintness and speedy effacement by carbon and 
hydrogen impurities, no fixed system can be recognised among them. The relative bright- 
nesses of the four recurring linelets a, b, b', c, whose positions are tabulated, are generally about 
5, 3, 1, 3 ; and it is the triplet of them a, b, c, whose positions, in metrical wave-lengths, are given 
for their recurring groups, in Angstrom's and Thalen's Table. 

In the present Table's columns of the linelet a, the metrical wave-length and wave- 
number are followed by the wave-length and wave-number in a British inch, so as to facilitate 

* Professors Liveing and Dewah's, and Dr Hugoixs' simultaneous recognitions of the remarkable ultra-violet 
spectrum of aqueous vapour in the light of all hydrogen-hearing flames (' Proceedings of the Royal Society of London,' 
June 1880), although announced just previously to the presentation of the above reflections, had not yet been received. 
Put they afford as yet no certain evidence that the same spectrum, indicative of aqueous vapour, is also producible by 
electrical dischargee in gas-vacuum tubes. 

•f- (a, c, See the accompanying sketch, p. 157). 



END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 155 

direct comparison of the new measures in the latter kind of wave-numbers (entered in the 
fourth and translated into metrical wave-lengths in the first column) with the chief linelet's 
wave-positions actually recorded in the Swedish observers' Memoir. The metrical wave- 
lengths of Angstrom's Table have also been translated accurately into British wave-numbers ; 
but they are only preserved in their original form (for the purpose of admitting a direct com- 
parison) in the case of the leading-lines a of the triplet groups. Among the other triplet- 
lines, comparisons of metrical wave-lengths are for brevity omitted, and only the equivalent 
values are retained in wave-numbers to an inch. Throughout all the comparisons presented in 
the Table, Angstrom's and Thalen's measures in their metrical or British values occupy the 
upper line in the space accorded to the position-observations of the several doubly-measured 
groups. The next, or middle line in duplicated groups, gives the measurement with the aurora- 
spectroscope; and under both are written average values of the two independently observed 
positions. Finally, the interval of the average value of each subordinate line's wave-number 
from that of its leading triplet-line, is inserted in the next following column contiguous to it, 
so that the regularity or degree of variation of these intervals in the successive groups and 
sections of the spectrum can be apprehended at a glance. The last of the intervals, on the 
right hand side of the Table, denotes the increase of wave-number from the first line of one 
linear group to the first line of that following it (repeated in position for this purpose in 
the next preceding column), or denotes the distance in wave-numbers between the first lines of 
successive groups. 

A mere inspection of the Table shows that while all the other intervals which it presents 
are sensibly invariable, or practically constant, the latter interval between the leading lines of 
successive groups is a constantly decreasing one. There are large and marked steps of this 
decrease at the 4th and 16th groups of the Table, followed by a rapid fall for one or two groups, 
and then a nearly constant interval for a long period of six or ten groups afterwards. In the 
remaining half of the flutings mapped by Angstrom and Thalen, but not now remeasured as 
they merge into the carbon-citron band, the same phenomenon is presented. At one or two 
groups after the carbon intrusion, the group interval falls abruptly from about 360 to 300, and 
remains at the average value of 295 for the remaining fifteen or sixteen groups to the green end 
of the series. Throughout the whole range of about forty groups mapped by Angstrom, the 
interval ac, nevertheless preserves a constant average value of 163 English wave-numbers 
between the two peaks of the double-notched serration. As far as Angstrom's and Thalen's 
groups are remeasured in the present Table, the common averages of the subordinate intervals 
of them are placed at the foot of their columns; and it appears that the line-intervals ab, ab' on 
the first slope have an arithmetical progression from the edge of the slope, or a common interval 
from line to line of about 52'3. The interval ac is perhaps not conformable to the progression, 
as although the tendency in measuring a bright edge is to place the pointer rather far upon it, 
and the excessive value of the interval ac= 163 - 5, shown in the average of the Table, might in 
this way possibly be accounted for, yet a different account of the discordance may perhaps be 
given from the following consideration : while in the remaining eight or ten complete ranks of 
Angstrom's list (following those here remeasured) the interval etc has a pretty constant value, 
as before, of about 162, the interval ab, though very uniform, has an average value of only 42*5, 
and this is not more conformable to the space of 162 or 163, than the former common difference 
of 52 - 3; while it differs entirely from that regular interval in the earlier portion of the 
spectrum. 

It appears probable, therefore from this review, that the two notch-edges of the nitrogen 



156 



TROFESSOR A. S. HERSCHEL ON 



WAVE-LENGTHS of Nitrogen Linelets between the extreme Red and Yellow;^) measured in a Cyanogen, and 
in a Deutoxide of Nitrogen End-on Vacuum Tube, with Prisms of the Aurora-Spectroscope giving a Dis- 
persion from A to H of 22° ; compared with Measurements of the same Lines by Angstrom and Tiial^n.. 
(References are made by numerals in the Table, to the brief list of Notes appended to it at the end.) 



1(?) 

2(?> 


a 


b 


(ab) 


V 


{ab') 


c 


(ac) 


a 
33,820 


(au) 

1150 
( = 2x 

575?) 


Appearances; and standard- 
line adopted places, in 
metrical wave-lengths, and 
in British wave-numbers. 


Wave- Wave- 
length ; number 
tenth- In ■ 
initios. mil-met. 


Wave- 
length ; 

British 

Inch. 


1 Wnve- 

nntnber in 
a British 

inch. 


7774-0 


1286-23 


o-ooo, 

030,609 


32,670 


( An air- 
l line; C.P.S. 
( (? oxygen. )(») 

I Koj, 7700 
( Ka,, 7069 


= 32,987 
= 33,140 


1 


7510-3 


1331-51 


29,568 


33,820 










34,003 


183 


34,355 


535 


( First two (ac) 
i of a bright, 
-{ sharp -lined 
| quartet (a c, 
I andc"d?).( 3 ) 




2 


7393-3 


1352-57 


29,108 


34,355 


34,411 


55 


34,469 


114 


34,530 


181 


34,911 


556 






3 


7275-6 


1374-46 


28,644 


34,911 










35,085 


174 


35,403 


552 






4 


7162-3 


1396-19 


28,198 


35,463 






35,573 


110 






35,950 ? 


487 






5 


7065-3? 


1415-37? 


27,816 


35,950? 










30,122 


172 


30,451 


501 


1 . 




6 


6968-2 


1435-09 


27,434 


36,451 






36,559 


108 


30,639 


188 


30,918 


407 


- 3 






[6870-0( 4 ) 
6880-1 

6785-7 
6791-6 


1455-60 
1453-48 

1473-69 
1472-41 


27,087 

26,715 
26,738 


36,918 

37,432 

37,400 
416 


36,972?] 
30,959 
965 

37,471 

37,463 

467 


47? 
51 


37,015 
37,522 


97 


37,090 

37,574 

(37,001?) 
587 


178 


37,416 
37,900 


498 
484 


[ Lia, 6705-5 


= 37,879 


106 


171 


'1 


6701-0 
6702 8 


1492-31 
1491-92 


26,382 
26.3S9 


37,905 

37,895 

900 


37,948 

37,955 

951 


51 


38,002? 


102? 


38,001 

38,081 

071 


171 


38,362 


402 






"1 


6621-8 
6620-2 


1510-16 
1510-54 


26,070 
26,064 


38,358 

38,367 

362 


38,403 

38,435 

419 


57 


38,477 


115 


38,510 

38,537 

526 


164 


38,827 


405 


[ Ha, 6562 


= 38,707 


H 


6542-3 
6541-0 


1528-48 
1523-82 


25,757 
25,753 


38,824 
38,831 

827 


38,875 

38,897 

886 


59 


38,931 


104 


38,979 

38,989 

984 


157 


39,282 


455 






H 


64655 
6466-5 


1546-67 
1546-43 


25,455 
25,459 


39,285 

39,279 

282 


39,327 

39,335 

331 


49 


39,380 


104 


39,437 

39,445 

441 


159 


39,734 


452 


f Oxygen(faint) 
( 0456-5 


| =39,340 


-1 




1564-38 

1564-37 


25,167 
25,107 


39,734 

39,735 

734 


39,782 

39,794 

788 


54 


39,831 


97 


39,894 

39,902 

898 


164 


40,182 


448 






»| 


6321 -0 
6321-8 


1582-03 
1581-95 


24,886 
24,887 


40,183 

40,181 

182 


40,229 

40,240 

234 


52 


40,292 


110 


40,350 

40,343 

346 


164 


40,641 


459 


1 Tube-C. 
"a" (faint red) 
§ 0298-5 


[ =40,327 


-1 


6249-2 
6250-5 


1000-20 
1599-89 


24,603 
24,608 


40,045 

40,637 

641 


40,688 

40,694 

691 


50 


40,737-66 
751 


110 


40,800 

40,800 

800 


159 


41,075 


434 






| 6188-2 

16 j 61843 


1617-29 
1617*02 


24,343 
24,347 


41,079 

41,072 

075 


41,133 

41,137 

185 


00 


41,152-204 

178 


103 


41,246 

41,253 

249 


174 


41,470 


395 


^ O (strong) 
S 0157 


J =41,254 


\ 6125-4 
17] 6124-4 


1632-55 
1682-82 


24,115 
24,111 


41,467 

41,474 

470 


41,511 

41,527 
519 


49 


41,573 


103 


41,625 

41,042 

633 


103 


41,871 


401 


"a 

S Li 0, 0102 

J 


= 41,025 










< rry forward (mini 


Mill 1 S 


tins of inter 


vals, (10)532 (] 


0)1054 (1 


0)1646 





, 



END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 



157 



Br 


ought forward (numbers and) 


sums of intervals, (10)532 


(10)1054 


(10)1646 




H 


a 


6 


(aft) 


ft' 
41,976 


(aft') 

105 


c 

42,030 

42,026 

028 


(«c) 
157 


a 
42,255 


(are) 
384 


Appearances ; and standard- 
line adopted places, in 
metrical wave-lengths, and 
in British wave-numbers. 


Wave- 
length ; 

tenth- 
metres. 


Wave- 
number 

in a 
mil-met. 


Wave- 
length ; 
British 
inch. 


\ Wave- 
number in 
a British 
inch. 


6066-3 
6066-3 


1648-48 
1648-48 


o-ooo, 

023,883 
23,883 


41,871 

41,871 

871 


( Tube-C. 
\ (str. orange) 
( 6078 


( =41,789 


"I 


6011-8 
6010-3 


1663-40 
1663-79 


23,669 
23,663 


42,250 

42,260 

255 


42,301 

42,309 

305 


50 






42,419 

42,420 

419 


164 


42,638 


383 






20 j 


5957-3 
5956-9 


1678-64 
1678-76 


23,454 
23,452 


42,637 

42,640 

638 


42,685 

42,694 

689 


51 


42,741 


103 


42,809 

42,801 

805 


167 


43,018 


380 






21 j 


5904-6 
5904-2 


1693-59 
1693-71 


23,247 
23,245 


43,017 

43,020 

018 


43,069 

43,073 

071 


53 


43,099-133 
116 


98 


43,179 

43,184 

181 


163 


43,400 


382 


) Naoj, 5895 
I Naa 2 , 5889 


= 43,086 
= 43,130 


22 j 


5853-0 

5852-2 


1708-53 
1708-77 


23,043 
23,040 


43,397 

43,403 

400 


43,448 

43,440 

444 


44 


43,494-527 
510 


110 


43,564 

43,570 

567 


167 


43,780 


380 


1 




23 


5801-8 
5801-5 


1723-60 
1723-68 


22,842 
22,841 


43,779 

43,781 

780 


43,829 

43,840 

834 


54 


43,871-905 
888 


108 


43,940 

43,946 

943 


163 


44,158 


378 


s Hydrogen 
-.§ (faint) 
g 5812-5 


• =43,698 


24 


5752-0 
5752-1 


1738-53 
1738-49 


22,645 
22,646 


44,159 

44,158 

158 


44,208 

44,209 

208 


50 


44,244-283 
263 


105 


44,323 

44,325 

324 


166 


44,531 


373 


J 




25 ] 


5703-8 
5704-1 


1753-22 
1753-14 


22,456 
22,457 


44,532 

44,530 

531 


44,577 


46 


44,610-643 
626 


95 


44,699 

44,678 
688 


157 


44,888 


357 






26 j 

■1 


5657-9 
5659-0 

5612-6 
5613-7 


1767-44 
1767-89 

1781-71 
1781-36 


22,275 
22,280 

•22,097 
22,109- 
-22,094 


44,893 

44,884 

888 

45,255 
45,231-61 
250 


44,950 


62 


44,986 


98 


45,058 

45,030 

044 

45,404 


156 
154 


45,250 
45,619 


362 
369 


) Gas-fl. Citr. 
\ line 1, 
) 5633-5 

Tube-C. 

Citr. edge, 

5608 


| =45,086 
= 45,289 






Total (numbers am 


i) averages 


of intervals 


(18) 


52-33 


(18 


)104-22 


(19 


163-47 






( 1 ) Double-notched band of a Nitrogen Serration. — The dotted lines 
show how Ha and Naa, fall among the linelets of the ridged band in which 
each of them occurs separately, in the red and yellow portions respectively 
of the channelled field. Only measurements of the linelets a, ft, ft', c of 
the flutings included in its range, are recorded in this Table. 

( 2 ) An extreme red ray, possibly a new oxygen line (?), seen with prisms of moderate dispersion in an air-vacuum tube ; as observed and 
measured in a paper on " End-on Vision in Private Spectroscopy " by Professor Piazzi Smyth. 

( 3 ) A fine close-membered cluster, disclosed by end-on vision at the extreme-red end of the spectrum in both of the Cyanogen and Nitro»en- 
deutoxide tubes. It consisted, when best exhibited and observed in them, of four not far from equidistant, about equally bright and exceedingly 
sharp lines. In one of the two measurements obtained of its positions before it lost its brightness in the tubes, it only showed three lines, by 
the loss apparently (as comparison with the four-line measurement seems to indicate) of its leading line. It forms at the red end of the spectrum a 
grandly protruding and'detached, somewhat distinctly formed, first linelet-group of the uninterrupted train of them there springing up. Its followers, 
or after-groups, though weak at first, soon brighten up into the long row of close flutings of the red-to-green Nitrogen serration. The observed 
wave-numbers of its third and fourth lines, 34,134 and 34,237 to a British inch (or wave-lengths 7441 '2 and 7423 - l tenth-metres) appear to a»ree 
best (although not quite precisely) with the average positions in the other groups, of the linelets marked in the above figure by the ridge-tops 
c"andd(?) 

( 4 ) This is the first spectral line noted in Angstrom and ThaliSn's Table. Although recorded there as a "first line," (re), its measure in 
English wave-number is here ranged under the same group's second line ft, with whose remeasured place it more nearly corresponds. The 
secondary pair 66', in fact, of this first-noted fluting (the next before that which comprises Lia, and the next but two before that in which Ho is 
included), a little outshines the leading line a, by blending together into a stronger blurred maximum of brightness at the beginning of the group. 

I o The varying intensities of the flutings are noted in the column of " Appearances " as "maximum " and "minimum," as they are given by 
Angstrom and Thal£n, and as they present themselves in the spectrum. Of the standard-lines noted in the same column, only Ha (faintly), 
and edge-places of the three Tube-carbon orange and citron bands (strongly visible) were measured in the Nitrogen tubes themselves, comparison- 
spectra being used to show the places and to furnish micrometer-readings of the other standard lines entered in that column. 



158 PROFESSOR A. S. HERSCHEL ON 

Uutings are really independent from each other in their derivations, and that the array of lines 
on their two slopes are not fellow-representatives of a common arithmetical progression, The 
general want of conformity among the line-intervals of the second slope may perhaps arise, 
accordingly, from the superposition upon each other of the two unconformable line systems of 
the two interfering slopes ; and a character which seems to be essential to the linelets of the 
second slope independently of every influence of conflicting impurities in the tube upon their 
comparatively slender strengths, possibly receives from a conjecture of this kind a satisfactory 
interpretation. That ruled and lined bands like those of olefiant-gas and other " carbon "- 
spectra possess, it would seem, insular characters in a spectrum, the minuteness of whose 
description, as revealed in the imposing tables of green gas-flame, and green tube-carbon bands 
drawn up by Professor Piazzi Smyth, surpasses comprehension, and almost registration, will I 
believe be granted from a close inspection of his observations and leductions. But if the 
inherent complexity of these shaded bands' internal structures proves to be so prodigious as the 
application of extraordinary dispersion shows, it is scarcely to be expected that among the close 
array of these ruled wedge-like luminosities crowded thickly into the nitrogen procession, order 
should reign among their leading, or frontier lines. Systems well studied on a larger scale 
appear here to be repeated, and innumerably multiplied in miniature. A comprehensive and 
far-reaching theory of banded and fluted spectra will therefore probably be required to include 
and account intelligibly for all the singular changes of orderly succession that the nitrogen 
flutings present when the positions of their leading edges, or in other words the space-intervals 
from spur, or terrace-edge to terrace-edge of the long serration, are further brought into com- 
parison with each other. 

In the extreme red, numerous groups were seen with the powerful end-on illumination, 
preceding any of those pictured and mapped by Plucker and Hittorff, and by Angstrom and 
Thalen. A few measures among these revealed a far-off line, star-like in its brightness, which 
proved to be a triple, and even quadruple line when brought into the middle of the field of 
view. It had already been measured and recorded accurately with low dispersion by Professor 
Piazzi Smyth as a Nitrogen-line ; and the result of the new measures of this part of the Nitrogen 
spectrum was to connect it, as shown in the Table, with a well-observed series of extreme-red 
nitrogen groups, of which it formed the first visible commencement. Its two least refrangible 
lines agree in their spectral positions with the two spurs or leading edges of a double-toothed 
serration, while its two more refrangible ones lie upon the fading flank of the second tooth's 
slope, at nearly the same equal distances asunder. In both Cyanogen and Nitrogen they were 
equally sharp and bright, like weak hydrogen or lithium lines, quite free from the haze and 
haziness connected with the corresponding linelets in other portions of the spectrum. 

But if end-on vision has revealed an outwork so substantial and remote as this of the dark 
red portion of the nitrogen procession, what may perhaps be gathered from the announcement 
of a " fine line " at the head of the list, in a view of the spectrum of air in an end-on vacuum 
tube, contained in a record of that spectrum as measured with a prism of ordinary dispersion, 
communicated by Professor Piazzi Smyth last year in his Paper on " End-on Illumination in 
Private Spectroscopy " to the Pioyal Scottish Society of Arts, the spectral position of which is 
1150 inch-units lower in its wave-number than this frontier line of nitrogen, and which is 
actually but little more refrangible than the dark-red Potassium-line itself ? * The interval to 

* In a careful search for low-temperature lines in an oxygen gas vacuum tube, Professor Piazzi Smyth has met 
with an extreme-red line at \V. No. 32,600 {circa), which may perhaps he identical with that above noticed as mapped 
in an "air-spectrum" at about W. No. 32,070. The latter line, in that case may perhaps b; an o.xygen-line, and not 
a nitrogendine as here supposed. 



END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 159 

this line, supposing its wave-number and that of the primary red nitrogen line to be very 
exactly fixed, is 1150 units, or two intervals of 575 inch-units each, which is about the length 
that the course of the intervals in the rest of the spectrum up to this unexplored portion, 
would lead us to expect. It may therefore be conjectured that the " air-line " noted in this 
place is a nitrogen-line, like the primary extreme-red one, of considerable brightness. The 
first object in the " air-spectrum " noted after it is a strong haze band occupying the place of 
the grand nitrogen leader, followed by two weaker haze bands in the places of the two 
next nitrogen groups; and finally in the extreme-red portion of the " air-spectrum" as seen and 
mapped with end-on vision by Professor Piazzi Smyth in the above mentioned Table, there are, 
from the first visible one, to the red hydrogen line, ten nitrogen haze-bands or serrations very 
well recorded in appearance and position, none of the places of at least half of which had ever 
been made sufficiently visible before for measurement, so as to afford useful data for instru- 
mental determinations and for theoretical discussions. 

The special capabilities of end-on illumination, for bringing under notice and exact 
observation an immense number of details not before iiivestigable or described, were 
exceedingly well displayed in this example. Data of the richest value, it cannot be doubted, 
are now being gathered, and views of the greatest insight and originality are in a fair way to 
be formed and fostered by the application of end-on vision and high optical perfection and 
dispersion to gas-spectroscopy ; but to have beheld the field of observation, and to have assisted 
the process as an admiring looker-on, has impressed me at the same time with the formidable 
as well as with some of the beautiful and splendid features of the scene ! It may be hoped 
that photography, in its now greatly improved practice both as regards general sensibility, 
and especially in that sensibility which relates to the red end of the spectrum, will ere long 
come to the spectroscopist's assistance, and relieve him of much the most serious and hindering 
portion of his labours, of disentangling and recording correctly what he sees. 

Now that it is well established that for a few lines sought to be produced and studied in 
a vacuum tube, a train of foreign lines and bands of contaminating gases commonly muster in 
the field, and blurr and confuse the natural spectrum sought to be examined almost beyond 
recognition, and when we further reflect that the intended spectra, if they are obtained in 
sufficient strength and purity to endure very great dispersion, are found to be of such exceeding 
intricacy as even then to surpass the means of accurate description, it will readily be admitted 
that photography would supply the spectroscopist most effectually with records of many little 
particulars of spectra containing important elucidations for his purpose, which pressure on his 
space and time constrains him, whatever art and skill he bestows upon their registration, 
to pass over unrecorded. Some sources of the irregularities of the nitrogen-spectrum-places 
were thus, it was thought, recognised in these observations, depending apparently on carbon 
and hydrogen contaminations ; but they are too long and various in their nature to be here 
narrated. 

„„ A common spectacle, however, both in the cyanogen and nitric-oxide tubes was a duplica- 
tion of the linelet b' (noted by the mean of the line-pair's places in the Table), and occurrences 
of extra linelets, sometimes between a and b, but more frequently on the fading-off declivity 
beyond c of the second down-slope of the double-notch. Although no doubt instructive, 
discussions of these diminutive characters of the spectrum, denoting perhaps only deficiencies 
of its strength and purity, would lead to very long and laborious descriptions. A list of 
standard line places, including three tube-carbon band-edges (red, orange, and citron) are there- 
fore added to the Table, to assist in distinguishing the groups, and to supply, in case of more 



160 END-ON VIEWS OF GAS SPECTRA UNDER HIGH DISPERSION. 

exact determinations being hereafter attempted of them, guides and possible explanations of 
some of their major variations. The tracts of dimness and brightness noted in the Table, which 
the range of groups present when they are most perfectly developed, are borrowed from 
Angstrom's and Thalen's Table in the above quoted Memoir, those observers, as well as 
I'lucker and Hittorff, in the admirable pictures which they have respectively published of it, 
having apparently mapped and noted the banded, or low-temperature spectrum of nitrogen, 
although not with all the structural details which it possesses, yet in its whole extent under 
the finest and most favourable conditions of completeness and perfection. 



ERRATA 

Page 103, line 13, erase " and double." 
Page 104, line 2, delete the second word. 
Page 143, top of 3rd column , for 37,707 read 38,707. 

On Plate XII., lowest spectrum strip, the single black line at 45 200 nearly, 
should be closer up to 4£ 000, or over the first line of the Citron band of the carbo- 
hydrogen group below, 

P. S. 



( 161 ) 



VI. — On a Special Class of Sturmians. By Professor Chrystal. 

(Read 20th June 1881.) 

If S„ be a rational integral function of x of the n th degree, and S„_i S„_ 2 . . . S x S 
a series of such functions of the n — l th , n — 2 th , &c, degrees, so related to S„ 
that, when any one of the whole series S S x . . . S„ vanishes, the two on 
opposite sides have opposite signs, and farther S„_ a and S„ have always opposite 
signs when x is just less than any real root of S„ — , then S S x . . . S re _i may 
be called a set of Sturmians to S n . It is obvious that the problem of finding 
such a set of functions admit of an infinite number of solutions. The first 
discovery of such a set was made by Sturm, and the researches of Sylvester, 
Hermite, and others have shown how other solutions of the problem may be 
obtained. 

It occurred to me while working at some physical questions that the 
properties of symmetrical determinants would furnish us with the means of 
constructing a particular class of Sturmians. I thought when I found the 
result that it was new, but a little research led me to a paper by Joachimsthal 
(Crelle's Journal, Bd. xlviii. p. 386), where the very same series is given. 
The method by which I independently arrived at the result is so simple and so 
different from that of Joachimsthal that I have thought it worth While to lay 
it before the Society. 



1. Let 



A = 



an «i2 

«21 #12 



a„i a n <i 






be a symmetrical determinant, so that a 12 = <% , &c. ; and let us call the deter- 
minant formed by deleting the first row and first column, the first two rows 
and the first two columns, and so on, its first, second, &c, principal minors. 

Then we have the well-known proposition that, if any principal minor 
vanish, the next higher and the next lower have opposite signs. This is easily 
vol. xxx. part i. 2 a 



162 



PEOFESSOE CHEYSTAL ON 



proved as follows : — Let A n A 12 , &c, the first minors corresponding to 
a u a l2 , &c, then (see Salmon, " Higher Algebra," p. 29) we have 



Hence, if 



Aji A 2 2 -A- 12 — (^11 • • • Unn)\ttz3 • • • drm) 

A n , i.e. (a 22 . . . a nn ), vanish, 



i.e., the determinant and its second principal minor must have opposite signs ; 
and similarly for the other cases, since all the principal minors are symmetrical 
determinants. 

2. In the next place, we have by the multiplication of matrices, 

%(a l a 2 . . . tyYZioi. 02 . . . a r )(x— a^)(x — a^) . . . (x — a r ) 



= (-1)'' 



:(-!)' 



11.. 


. 1 


X 


1 








x a x . . 


• a n 




of 


• a/ 


x 2 a 2 . . 


■ a n 2 




a 1 ?+ 1 . . 


«/ +1 


x r a{ . . 


• a/ 




a 1 ^+'- 1 . . 


0/+'- 1 


1 s p 


. . Sp + ,._i 


= S r (x) say, 




X Sp+1 ' 


. . s p+ 


** 









X" Sp + 2 



i// Op-i-f 



Sp+r+l 



s p+2r-l 



Here a : . . . a n are any n quantities real or imaginary ; £(04 a 2 . . . a r ) denotes, 
according to Sylvester's notation, the product of the squares of all possible 
differences of c^ a 2 . . . a r ; and 2 denotes summation with reference to all 
possible groups r at a time of the n quantities. 

It is obvious that, if r>n, then S r (#) = ; and if r = n, 



s„c*o=(-i)" 



Sp+l Sp +2 



. x n 

■ Sp+n 



Sp+n — 1 Sj,+ n Sp+n+1 • • • Sp+2n-l 

= {a l a. i . . . a,^(a x . . . a„)(x — a i )(x—a 2 ) . . . (x—a n ) 

= (0^2 . . . a„)''£(«i . . . a >l )(x n +p 1 x n -' i +p 2 z n - 2 + . . . +p„) 



A SPECIAL CLASS OF STURMIANS. 
if 04 a 2 . . . a re be the roots of 

x n +p 1 x n ~ 1 + . . . +p n = 0. 
3. When % = a 1 
dS„(x) 
and 



163 



dx 



becomes (a x a 2 . . . a n ) p ^(ai . . . a n )(a 1 — a?) . . . (a r — a,i) 



S«_i(aj) becomes (a 2 a 3 . . . a n ) p %(a 2 . . . a n )(a 1 — a 2 ) . . . (a t — On) 

Hence, when x = a 1 



dS„(x) 
dx / 



S n -i(x) = a 1 P{(ai — a 2 )(a l — a 3 ) . . . (c^ — a„)] 2 . 



Now, if a x be real, and p be an even positive or negative integer, this ratio will 
be real and positive ; for a x a 2 . . . a„ being by supposition the root of an equa- 
tion with real coefficients, for every imaginary in the series a x — a 2 , a : — a 3 . . .'a x — a n , 
there will occur a corresponding conjugate imaginary so that the product of them 
all will be real. 

It follows that S„_i(#) and S„(#) have opposite signs when x is just less than 
any real root of 

S n (#) = 0, 

which is the second characteristic of the first two functions of a Sturmian 
series. 

The restriction as to p being even may be removed if positive and negative 
roots be considered separately ; but for simplicity I shall suppose p to be 
always even. 



4. If we take the determinantal expression for S„ , multiply each column 
by x, and subtract the next following, leaving of course the last column 
unchanged, we get, denoting for brevity s p x— s p+1 by (p) , s p+1 x—s p+2 by 
(p + l),&c, 



S n (x) = 



(P) + 1)0 + 2) . . . (p+n-1) 

(p + 1) (p + 2)(p + 3) . . . (p+n) 
(p + 2) (p + 3)(p+4) . . . (p+n + 1) 

(p + n — l)(p + n)(p + n + l) . . . (p-\-2n — 2) 

which it will be observed is a symmetrical determinant. S„_i(#) , similarly trans- 
formed, becomes the first principal minor of this obtained by deleting the last row 
and the last column, and so on. Hence, by (1) , S n (x) , S„_i (x) . . . S x (x) S (x), 



1G4 



PROFESSOR CHRYSTAL ON 



the last being any positive constant, have the property that, when any one 
of the series vanishes, the next higher and the next lower have opposite 
signs. 

5. It has now been shown that S n (x) , S n _i(x) , . . . S^cc) , S (x) form a 
Sturmian series. By giving particular even values to p, we get of course an 
infinite number of such series. 

If it were desirable to employ these functions for the purposes of root 
discrimination, s p , s p _ lt &c, could be calculated by Newton's method, and by 
giving a proper negative value to p, the labour could be diminished by nearly 
half in the most general case. 

For example, if we take the cubic equation 

x 3 +px + q = , 



and put p = — 2 , the Sturmian's are 



S s = 



1 X X 2 X 3 


; °2 — 


+ 


1 X X 2 


> s x = — 


1 X 


S_ 2 S_i S S x 






S_ 2 S_l S 




S_ 2 S_i 


S_i S Si $2 






S-l s s x 




S-l s 


Sq Si S2 S3 






So Si s 2 







S =+l. 



6. If we wish simply to find how many real roots there are, then we have 
simply to consider the signs of the coefficients of the highest powers of x in 
the Sturmians. This gives us the following theorem : — 

There are as many pairs of imaginary roots of the equation 

x n +2h xn ' 1 + • • • +Pn = 
as there are variations of sign in the series 



■fJ-; Sp , 



Sp Sp+i 
Sp+1 Sp+2 



Sp Sp+l Sp+2 
Sp+1 Sp+2 Sp+3 
Sp+2 Sp+3 Sp+4 



&c. 



when p = this gives a well-known theorem (see Salmon, " Higher Algebra," 
p. 49). 

If we put p = , the series for the cubic 

x 3 +px + q = , 

neglecting certain positive multipliers, is 

+ 1, +3, -6p, -(4p* + 27q 2 ). 



A SPECIAL CLASS OP STUPMIANS. 165 

If we put p — — 2, we get 

+ 1, +p 2 , + 2p 2 , -(4p 3 + 27^ 2 ). 

Each of these leads to the well-known condition for the reality of the roots 
of the cubic. 

7. It follows at once from (2) that, if two roots of the equation be equal, 
then S„(#) vanishes identically, and S^^a?) , S„_ 2 (#), . . . S (#), form a Stur- 
mian series for the roots all supposed single. If three roots be equal to one 
another, or if two pairs be equal, then S„(x) and S n _ 1 («) vanish identically, and 
the rest form a Sturmian series for all the roots supposed single ; and so on. 
The present class of Sturmians present therefore an instructive contrast to the 
ordinary series obtained by the method of the greatest common measure. 



VOL. XXX. PART I. 2B 



( 167 ) 



VII. — On the Cranial Osteology of Rhizodopsis. By Ramsay H. Traquair, 
M.D., F.R.S., Keeper of the Natural History Collections in the Museum 
of Science and Art, Edinburgh. 

(Read May 21, 1877. Eeceived for Press July 22, 1881. Abstract in "Proceedings," vol. ix. p. 444.) 

In a paper by Mr E. W. Binney on the Fossil Fishes of the Pendleton Coal 
Field, published in 1841, the dentary bone of Rhizodopsis is figured as the 
" upper jaw of a new species of Holoptychius," to which, however, he did not 
attach any specific name. In the same paper its scales are also figured and 
referred to the same genus.* Scales belonging to the same fish were after- 
wards figured by Professor Williamson under the name of Holoptychius 
sauroides,^ and again by Mr Salter, as those of Rhizodus granulatus.% Both 
of these specific names occur under Holoptychius in Acassiz's general list of 
Ganoids published in 1843, but as they were unaccompanied either by figures 
or descriptions, it is really immaterial which of them, if indeed either, was 
applied by him to the fish in question. The authority for the term "sauroides" 
as applied to the common species of Rhizodopsis, the only species of the genus 
which is as yet known with certainty, must therefore remain with Professor 
Williamson. Holoptychius sauroides of Binney § and of Messrs Kirkby and 
Atthey || is quite another fish, now also distinguished generically as Strepsodus, 
and for it the specific name " sauroides" is therefore equally valid. 

In 1866 Professor Young published a description of the entire fish, under 
the name of Rhizodopsis sauroides, Williamson, sp., the authorship of the new 
generic title being attributed to Professor Huxley. If From Professor Young's 
description, we learn that the position of Rhizodopsis, in Professor Huxley's 
classification of the Ganoids, is in the cycliferous division of the Glyptodipterine 
family of the suborder Crossopterygidse, and that it possesses subacutely lobate 
pectoral fins, two dorsals, and a heterocercal tail. Some of the bones of the 
head are noticed, such as the parietals, the three dermal plates of the occipital 
region, the opercular bones, the maxilla, and the mandible. No prsemaxilla 

* Trans. Geol. Soc. Manchester, vol. i. (1841), pp. 153-178, pi. v. figs. 6, 8, and 10. 

+ " On the Microscopic Structure of the Scales and Dermal Teeth of some Ganoid and Placoid Fish,' 
Phil. Trans., 1849, p. 457, pi. xlii. figs. 21-23. 

\ "Iron Ores of Great Britain," Mem. Geol. Survey, 1861, p. 223, pi. i. figs. 4-6. 

§ Op. cit, pi. v. fig. 7. 

|| Trans. Tyneside Nat. Field Club, vol. vi. (1863-64), p. 234, pi. vi. figs. 5 and 6. 

IT "Notice of New Genera of Carboniferous Glyptodipterines," Quart. Journ. Geol. Soc, 1866, 
pp. 596-598. 

VOL. XXX. PART I. 2 C 



168 RAMSAY H. TRAQTTAJR ON 

was, however, observed by Professor Young, and he states that the jugular 
plates are " in two pairs, principal and posterior," and that there is no trace of 
median or lateral plates. The characters of the scales and of the vertebrae, 
whose centra are in the form of osseous rings, are described as well as the 
dentition ; the teeth of the maxilla being fine, equal, and conical, while those of 
the mandible are of two sizes. The non-trenchant character of the mandibular 
laniaries distinguishes the genus from Rhizodus, while as separating it from 
Holoptychius, Professor Young gives the thinness of the scales, the nature of 
their ornament, and the presence of teeth of two sizes. 

Two years later a notice of this fish was published by Messrs Hancock and 
Atthey, from specimens found in the shales of the Northumberland Coal Field,"* 
in which the authors state that in all respects their specimens " agree well with 
Dr Young's description of the species." Their description contains, however, 
two points specially worthy of notice, viz., the detection, on the anterior 
margins of some of the fins, of peculiar fulcral scales similar to those which 
occur in Megalichthys and other Saurodipterines, and the determination of a 
peculiarly shaped dentigerous bone as " prcemaxilla." Moreover, according to 
Messrs Hancock and Atthey, the piscine genera and species Dittodus parallelus, 
Ganolodus Craggesii, and Characodus confertus, and the supposed Amphibian 
Gastrodus, all founded by Professor Owen on specimens of teeth from the same 
coal-field, are only synonyms of Rhizodopsis sauroides. 

Rhizodopsis is also noticed by Mr T. P. BARKAS,t who accepts Messrs 
Hancock and Atthey's interpretation of the bone supposed by them to be a 
prgemaxilla. So also does Mr J. W. Barkas,J who solves the problem 
regarding the specific nomenclature of the fish by quoting Rhizodopsis sauroides 
and granulatus as distinct species, without, however, giving any reasons in 
support of the supposed distinction. 

Being struck by the total dissimilarity of form presented by the bone 
interpreted by Messrs Hancock and Atthey as the prasmaxilla of Rhizodopsis, 
when compared with that element in other Crossopterygii, I carefully examined 
the subject with the aid of a beautiful series of specimens from North Stafford- 
shire, kindly lent me by my friend Mr John Ward, F.G.S., and with the result 
of finding that the reputed prasmaxilla is in reality the dentary element of the 
mandible. Moreover, the mandible of Rhizodopsis is of a very complex 
structure, and that structure finds itself in all essential respects repeated and 
explained in the mandible of the much more bulky Rhizodus Hibberti. 
These observations were published in the " Annals and Magazine of Natural 

* " Note on the Remains of some Reptiles and Fishes from the Shales of the Northumberland 
Coal Field," Ann. Nat. Hist. (4), vol. i. (1868), pp. 346-378. 

f " Manual of Coal Measure Palaeontology," London, 1873, pp. 23-25, Atlas, figs. 59-66. 
X Monthly Review of Dental Surgery, vol. iv. No. x., March 1876. 



THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 



169 



History" for April of the present year (1877). In the present communication I 
propose, with the aid of a few restored outline drawings, to consider the entire 
subject of the cranial osteology of Rhizodopsis, the greater part of the material 
for which belongs to the collection of Mr Ward. My thanks are also clue to 
Mr John Plant of Salford, for the loan of a number of shale specimens, showing 
isolated bones, from the Manchester coal-field. 



Rhizodopsis sauroides, Williamson, sp. 

Cranium proper. — The cranial roof bones form a " buckler," which in its 
configuration and composition is very similar to that in Osteolepis, Megalichthys, 
&c. As in these forms it falls into two principal parts, anterior and posterior, 
of which the posterior, or parietal portion, is slightly longer than the anterior 
or fronto-ethmoidal. The parietal portion is 
about twice as broad posteriorly as it is in front, 
each external margin passing, a little behind 
the middle, first inwards at an obtuse angle and 
then nearly straight forwards ; the anterior and 
posterior margins are nearly straight. This 
portion of the buckler is composed of six paired 
ossifications, two of which {pa. fig. 1) extend 
along its whole length, articulating with each 
other in the middle line ; their form is rather 
narrow and elongated, and they are also broader 
behind than in front. These two plates may 
very safely be reckoned as the parietals ; as 
such the corresponding plates have been, in 
Osteolepis and Megalichthys, designated by Pan- 
der, by Huxley in Glyptolcemus, and by Agassiz 
in Osteolepis, although the last-named author has 
marked the very same bones in Megalichthys 
as "frontals." Along the outer edge of each parietal are two smaller plates, 
anterior (/?/.) and posterior (sq.), regarding the signification of which, in 
allied forms, some pretty serious difference of opinion is found in the works 
of different writers. By Agassiz the anterior one was, in Osteolepis, con- 
sidered to be the post-frontal, the posterior to be the "mastoid," while 
in Megalichthys, he considered the very same plates to be equivalent to 
the chain of intercalary ossicles placed along the external margins of 
the cranial shield in Polypterus. By Pander the latter interpretation is 
accepted both for Osteolepis and Megalichthys ; while by Professor Huxley, 
these two plates, anterior and posterior, are in Glyptolcemus respectively termed 




Fig. 1. — Upper Surface of the Head of 
Rhizodopsis sauroides. 
s. t. supratemporal ; pa. parietal ; sq. squa- 
mosal ; p.f. posterior frontal ;/. frontal; 
or. orbit ; p.mx. prsemaxilla. 



170 RAMSAY H. TRAQUAIR ON 

" post-frontal " and " squamosal." Now, as the bones of the skull of Teleostean 
fishes, known in the Cuvierian system of nomenclature as " post-frontal " and 
" mastoid," are ossifications in the periotic portion of the primoidial cranium 
{sphenotic and pterotic of Parker), and as the disputed bones in the cranial 
buckler of the Crossopterygian Ganoids above referred to are evidently dermal 
in their nature, the latter may be considered as really partaking more of 
the nature of the ossa intercalaria in Polypterus. But as to their being 
considered exactly the equivalents of those little plates in Polypterus, there are 
some pretty serious, and to my mind fatal objections. They are firmly united 
by suture to the outer margin of each parietal, with which they form an integral 
part of the cranial buckler. In the Lepidosteoid Ganoids {Lepidosteus, 
Lepidotus, &c), there is, external to each parietal, a plate [squamosal) evidently 
corresponding to the posterior of the two in Rhizodopsis, &c, and which no 
one has ever thought of considering homologous with the Polypterine inter- 
calaries. The same plate is found in Amia, and there is in addition another 
smaller one in front of it corresponding to the anterior of the two in Rhizodopsis, 
but which, from the relatively greater shortness of the parietal, and the corre- 
sponding greater extension backwards of the frontal, comes to lie external to 
the posterior part of the outer margin of the latter. In the Palaeoniscidse there 
are also two corresponding plates, but the anterior of these, which I have 
lettered as post-frontal in my memoir on the structure of this family,* is placed 
relatively to the frontal still further forwards, owing to the greater proportional 
length of the squamosal behind it. In Acipenser there is also, external to 
the plates which seem to represent the parietals and frontals of other fishes, a 
chain of two or more smaller plates, which apparently represent those in 
question, and which, firmly articulated with the others covering the cranial 
cartilage, lie inside the position of the spiracle. There is no spiracle in 
Lepidosteus or Amia, and no evidence of it in the Palseoniscidas, or in either the 
Rhombo- or Cyclodipterine Crossopterygii, but in Polypterus there is, and the 
chain of intercalary ossicles, loosely articulated to the margin of the cranial shield, 
lies external to the spiracular slit, which passes down between two of them 
and the side of the cranium proper. It therefore seems to me inappropriate 
to consider the bones p.f. and sq. of the cranial shield of Rhizodopsis and allied 
forms to be the homologues of the intercalary ossicles in Polypterus, and better 
to follow Professor Huxley in designating them respectively as post-frontal and 
squamosal, always bearing in mind, however, that the former has nothing to 
do with the post-frontal of Cuvier, for which it is better to adopt the term 
" sphenotic " as proposed by Parker. In Amia, in fact, a well-developed 
sphenotic coexists with the more superficial plate to which I have referred as 
" post-frontal." 

* "Carboniferous Ganoids," Palrcontographical Society, 1877. 



THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 171 

The anterior, or fronto- ethmoidal division of the cranial shield is not so 
well preserved, so that it is not possible to map out its constituent ossifications 
with completeness ; in no case are its external or orbital margins well defined, 
and its upper surface is more or less broken and crushed. Nevertheless, the 
form and constitution of its anterior margin are unmistakeable. This is 
erescentically expanded, forming the rounded depressed snout ; and to the two 
dentigerous bones, the prcemaxillce forming its oral edge, we shall presently 
return in describing the bones of the jaws. I have not been able to detect the 
nasal openings. 

The external surfaces of these cranial plates are ornamented with minute 
tubercles and short ridges, frequently arranged in lines radiating from the 
centres of ossification. 

Facial Bones. — Immediately behind the posterior margin of the cranial 
shield are the usual three plates (s.t., fig. 1), one median and two lateral, which 
are of such constant occurrence in fishes of the Rhombo- and Cyclodipterine 
families. I have already, in my memoir on the structure of THstichopterus 
alatus* expressed my opinion that these are equivalent to the transverse chain 
of supra-tempoval ossicles in Polypterus, Lepidosteus, &c. 

The hyomandibidar is a somewhat elongated bone, extending downwards 
with a slightly backward inclination from below the squamosal to just behind 
the articulation of the lower jaw ; it is also slightly curved, the concavity being 
directed forwards. Above, where it articulates with the cranium, it is flattened 
for about a little less than one-third of its length ; this flattened portion, to 
which the superior anterior angle of the operculum is articulated, becomes very 
suddenly cut away on the posterior aspect, below which the bone becomes 
slender and cylindrical, expanding, however, in thickness in its lower half. 
Remains of a powerfully developed palato-quadrate apparatus are seen in several 
specimens, but not exposed with sufficient completeness to admit of any de- 
scription of its component elements ; its outer margin is for some distance 
articulated with the inner aspect of the maxilla, behind which it recedes a little 
inwards to admit of the passage of the masticatory muscles to the coronoid 
part of the lower jaw. 

By reason of the slightly backward slope of the hyomandibular, the gape is 
wide, and in three specimens, it is exposed all round the head, so that the 
bones forming the edges of the mouth are very completely seen. In nearly all 
the heads preserved in nodules the upper margin of the maxilla (ma:, fig. 2) is 
injured, but its complete contour is well exhibited in detached shale specimens. 
In shape it resembles very closely the maxilla of Megalichthys, being of an 
elongated triangular form, broadest about the junction of its posterior and 
middle thirds, and narrowly tapering anteriorly. Its posterior extremity forms 

* Trans. Roy. Soc. Edinburgh, vol. xxvii. (1874) p. 386. 



172 



RAMSAY H. TRAQUAIR ON 




Fig. 2. — Lateral View of the Head of Rhizodopsis sauroidcs. 
op. operculum; s.op. suboperculum ; p. op. prseoperculum ; 

x. x 1 . plates on the cheek; /. principal jugular; l.j. 

lateral jugular ; m.j. median jugular ; mx. maxilla ; 

d. dentary ; ag. angular; i.d. infradentary ; or. orbit; 

s. o. suborbital ; s. t. supratemporal ; pa. parietal ; sq. 

squamosal; p.f. posterior frontal; f. frontal; p.rnx. 

prpemaxilla. 



a tolerably acute angle, from which the inferior margin slopes first a little 
downwards and forwards, and then passes nearly straight forwards ; the short 
posterior margin slopes gently upwards and forwards to the very obtuse and 
usually more or less truncated superior angle, from which the superior margin 

then slopes downwards and forwards 
to the anterior extremity, just before 
attaining which it sends off a small 
articular process directed obliquely 
upwards and forwards. The external 
surface is ornamented with minute 
pits and delicate reticulating ridges ; 
the inner surface shows a delicate 
ledge running longitudinally a little 
above the inferior margin and nearly 
parallel with it. The inferior margin 
of the maxilla is set with a single row 
of small teeth, cylindro-conical, acutely 
pointed, slightly incurved, and of 
equal size. Their external surfaces 
are quite smooth and glistening under 
an ordinary lens ; they are usually placed pretty closely together, though some 
irregularity in their distances from each other is not unfrequently observed. 
Each of these teeth measures about 4 V inch from base to apex in a maxilla of 
1-L inch in length. 

In several specimens are seen the sharp imprints of two small dentigerous 
bones (p.mai.) forming the front edge of the mouth below the snout, and placed 
between and articulating with the anterior extremities of the right and left 
maxillae, while they are joined with each other in the middle line. Each of 
these two bones is nearly as high as long ; they are firmly fixed to each other, 
and also to the front of the cranial shield ; the posterior extremity of each fits 
into the angle between the anterior extremity of the maxilla and the little arti- 
cular process already mentioned in the description of the last-named bone ; the 
attached teeth, seen in impression and in section, resemble those of the maxilla. 
That we have here the true prcemaxillce cannot for a moment be doubted ; it 
is therefore abundantly clear that this element in Rhizodopsis does not in the 
least resemble the bone interpreted as such by Messrs Hancock and Atthey, 
but that on the other hand it is quite conformable to the type of prsemaxilla 
found in other Crossopterygii, as indeed in the Ganoids generally. 

The mandible is longer than both prasmaxilla and maxilla put together, 
reaching, as it does, a little further back than the posterior extremity of the 
latter. Its depth is contained about four times in its length, its upper and 



THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 173 

lower margins are tolerably parallel save just at the anterior extremity, where 
the upper one bulges a little upwards in a slight convexity, and at the posterior 
extremity where the same margin suddenly slopes downwards and backwards 
at an obtuse angle, meeting the lower one, which likewise curves upwards 
towards it, in a posteriorly directed point. Nothing has been said in the works 
of previous writers concerning the constitution of the mandible, though it 
might be inferred to be a composite structure, as it is in all fishes with ossified 
skeleton, and more especially in the Ganoiclei. In one specimen we find that 
over a considerable area the bony matter of the outer aspect has flaked off, 
leaving behind it a pretty sharp cast with sutural lines. On close examination 
a suture is seen commencing near the posterior extremity of the upper margin 
of the jaw, which, passing gradually downwards and forwards, marks off as 
dentary (d. fig. 2) an element precisely the counterpart in shape of the bone 
reckoned by Messrs Hancock and Atthey " prsemaxilla," but here placed with 
its toothed margin upwards instead of downwards as supposed by them. These 
two bones, right and left, are in many specimens indisputably seen forming the 
lower margin of the mouth and meeting each other at the symphysis. Each 
dentary bone is of a somewhat narrow and elongated form, truncated and some- 
what expanded at the anterior or symphyseal extremity, and pointed at the other 
or posterior. The upper margin, nearly straight, save just in front where it shows 
a slight convexity, is set with a single row of small pointed teeth of nearly uniform 
size, but the anterior extremity bears in addition a single more or less incurved 
laniary tooth, much larger than the others, and also more internal in its 
position ; the opposite margin, thin and sharp, displays a gently flexuous 
contour. Seen from the inner aspect, the anterior extremity of the bone 
presents a conspicuous thickening, in which the large laniary tooth is socketed, 
and which at the dental margin passes into a delicate ledge, which runs back 
for some distance along the roots of the smaller teeth. The teeth borne by 
this bone are round in transverse section, slender-conical in shape, brilliantly 
polished, and apparently smooth externally, but under a lens the surface is seen 
to be delicately fretted with minute longitudinal groovings, disappearing 
towards the point ; the large laniary is also very distinctly fluted or plicate at 
its base. 

The rest of the outer surface of the mandible is composed of at least three 
additional bony plates, separated from each other by sutures which pass 
obliquely forwards and upwards. The posterior and largest of these (ag. fig. 2) 
covering over the articular region, may be considered as equivalent to the 
angular element, though it also occupies very much the place of a supra- 
angular ; the other two (d.) in front of the latter and below the dentary, 
may be called infradentary . The presence and contour of these large infra- 
dentary plates is perfectly clear, the evidence as to additional ones is obscure. 



174 RAMSAY H. TRAQUAIR ON 

From the appearance presented by one specially large mandible, I rather 
suspect there is a third small one, as there is in Rhizodus, just below the 
symphyses! extremity of the dentary, and I have in my paper in the " Annals " 
referred to some doubtful evidence of still another, situated posteriorly on the 
lower margin of the jaw, and here separating the angular from the first infra- 
dentary for a little distance, but on this I am not prepared to insist. 

We have as yet accounted for the attachment of one laniary tooth, the one 
at the symphysis. But the mandible of Rhizodopsis , when perfect, shows not 
merely one large tooth in front, but several additional ones (usually three in 
number) behind it and internal to the series of smaller teeth. What has 
become of these in the dentary bone when disarticulated and detached ? 

A ready explanation of this is found in the structure of the lower jaw of 
certain Old Red Sandstone " Dendrodonts " in which the laniary teeth are not 
attached to the dentary bone proper, but to a series of accessory "internal 
dentary " pieces articulated to its inner side."" Should this also be the case 
with the posterior laniaries of the mandible of Rhizodopsis, then in cases where 
its elements are broken up and separated, these additional pieces will also get 
detached, and the absence of all but the anterior laniary in the isolated dentary 
bone will thus be amply accounted for. 

At the time I wrote the notice in the "Annals," already quoted, I had not 
obtained a clear view of the ossicles supporting the posterior laniaries in 
Rhizodopsis, and consequently referred to the analogy of the structure of the 
lower jaw in Rhizodus, in which I had most certainly found them, as amounting 
to a moral certainty of their existence also in the former genus. My attention 
has subsequently been directed to a specimen in the Edinburgh Museum of 
Science and Art, which completely confirms the view I then took. 

This is a slab of shale, not localitated, but probably from the Edinburgh 
Coal Field, over which scales of Rhizodopsis of large size lie thickly scattered, 
some of which are over 1 inch in length and nearly f in breadth. This is 
indeed an unusually large size, but is by no means an isolated example of the 
bulk which Rhizodopsis must sometimes have attained, and the form and 
sculpture of the scales here exhibited unmistakeably demonstrate the genus 
to which they belong. Lying in the midst of the scales is a mandible, 
evidently belonging to the same fish, and seen from the internal aspect. 
The splenial is gone, as is likewise the bony substance of the symphyseal part 
of the entire mandible, though a rough impression of it remains on the stone ; 
the hinder extremity is also injured, as well as the posterior part of the lower 
margin ; such impressions of the external surface, as remain when the bone has 
splintered off, indicate a sculpture of the usual minutely pitted-rugose character 

* See Pander's " Saurodipterinen, Dendrodonten, &c, des devonischen Systems," pp. 41-43, 
tab. x. figs. 2, 3, 4, 14, 22. 



THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 175 

of the mandibular elements of this genus. The depth of this jaw is 1 T ^ 
inch ; its entire length, including the impressions of its anterior and posterior 
extremities, is 5^ inches. The upper edge of the dentary element is seen 
extending from the obtuse angle of the posterior extremity of the upper aspect 
of the jaw to where it is broken off, apparently 1^ inch from its symphyseal 
termination, as indicated by the impression, and is set with a single row of small 
conical teeth, placed on an average at distances from each other of -^ inch, 
though they are more closely set anteriorly, where a few empty sockets are also 
seen. Some of the hinder ones are entire, and measure ^ inch in length ; they 
are sharp, slightly incurved, their bases plicate, the surface fretted with very 
minute strise, visible only under a strong lens. Anteriorly they are all broken 
off at various heights, the sections showing a large internal pulp cavity, the 
walls of which become very simply plicate at the base. Now, articulated just 
below this dentary margin is a longitudinal chain of two separate ossicles and 
the hinder part of a third. Each of these (int. d.) is of an oblong shape, con- 
tracted at the extremities, and in the middle showing first an empty socket, and, 
immediately in front of this, the broken off root of a large laniary tooth, at once 
recognisable by the complex folded structure of its constituent dentine. The 
anterior of these ossicles is obliquely broken off right through the empty 
socket, at the bottom of which are the remains of dentinal plicae, showing how 
here too a large tooth had once existed and had been broken off; and in front 
of this, and just above where the root of the actual laniary had been, is a part 
of the impression, upon the matrix, of the very tooth itself. Nothing can be 
more distinct than the sutures which separate these accessory or internal 
dentary ossicles from each other, and from the contiguous dentary element 
proper — the remaining bony matter beneath, consisting of the plates previously 
referred to as angular and infradentary, is thin and traversed by numerous 
cracks and fractures, so that very careful examination is here required for 
the determination of sutures. Nevertheless, with due attention, the lines 
of demarcation between the angular and the two large infradentaries may 
be made out, and just behind the position of the symphysis there is an 
indication of another suture passing upwards and forwards from the lower 
margin of the jaw, and separating off the third and smaller infradentary 
already alluded to. Lying on the margin of the slab, 2^ inches from the above- 
described jaw, is a broken-off piece of bone having a large tooth attached to 
it, the latter measuring § inch in length by i inch in diameter at the base. Its 
length was originally in all probability greater, as it is obliquely fractured, and 
the fractured surfaces ride over each other a little. Its base is plicate, above 
which the surface of the tooth is very minutely and delicately striated up to ^ 
inch from the point, which is perfectly smooth. Close beside this large tooth, 
and apparently attached to the same piece of bone, are two smaller ones, each 

VOL. XXX. PART I. 2D 



176 RAMSAY H. TRAQUAIR ON 

about I inch in length, so that I rather think we have here a fragment of the 
anterior extremity of the clentary bone of the other side of the head, with the 
symphyseal laniary. 

Returning to the examination of the smaller specimens, a portion of the 
splenial element is seen in one specimen, exposed by the breaking out of a 
portion of the middle of the mandible. The articular element, which was 
doubtless also present, is not exhibited in any specimen I have seen. 

The opercular bones are largely developed. The operculum {op. fig. 2) is a 
large, somewhat square-shaped plate, though broader above than below, and 
behind than in front. Its posterior-superior angle is rounded off ; its inferior 
margin overlaps another plate, which may be considered to be the suboperculum 
(s. op.). This is somewhat narrower, and has its posterior-inferior angle much 
rounded off; its upper and lower margins are nearly parallel, and from the 
former, just at the anterior-superior angle of the bone, there projects a short 
pointed process, producing the anterior margin a little way upwards. 

In front of the operculum, and covering a large part of the cheek, is a plate 
(x) of a somewhat oval shape, and somewhat obliquely placed, so that its long 
axis runs from below upwards and forwards. Above, it is in contact with the 
outer edge of the cranial shield ; its posterior margin is separated from the 
operculum by a smaller plate {p. op.). The latter is of a narrower shape, rather 
pointed above and a little less so below ; its long axis is pretty parallel to the 
direction of the hyomandibular which it covers ; its posterior margin, in contact 
with the operculum, is gently convex ; its anterior one, somewhat angulated, 
articulates with the large plate x, and below also with the smaller one x' . 
This third plate x' lies immediately above the articular extremity of the 
mandible ; its posterior margin, covering the lower extremity of the hyoman- 
dibular, is in contact with the suboperculum below, touching also the plate 
p. op. above ; its upper margin is articulated with the plate x, while in front it 
comes into relation with the oblique posterior margin of the maxilla. As 
figured by Agassiz, three precisely similar plates occur in the same position in 
Megalichthys* of which he compares both the upper and posterior to the 
so-called prge-operculum of Polypterus, while the lower one he compares to the 
little bone fixed above the posterior edge of the maxilla in the Salmonidse, &c, 
and which by Mr Parker is considered to be the homologue of the malar bone of 
other vertebrata.t In Osteolepis, according to Pander, the corresponding space 
on the cheek is occupied by one large plate, denominated by him "prseo- 
perculum," on which, however, lines are visible indicating a division into three 
similar component parts. On comjjaring the arrangement with what is seen in 

* " Poisson's Fossiles," vol. ii. part 2, p. 92 ; " Atlas," voL ii. pi. lxiii.a, figs. 1 and 3, i, k, I. 
\ " On the Structure and Development of the Skull in the Salmon " (Sahno salar, L.), Phil. Trans., 
1872, p. 100. 






THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 



177 



Polypterus, it is, I think, pretty evident that the hone p. op., together with the 
one x in Rhizodopsis, corresponds to the large cheek plate in the former genus, 
considered by Agassiz to consist of the equivalents of the cheek cuirass in 
Lepidosteus united with the prseoperculum, while the lower one (x') apparently 
corresponds to the posterior of the two small plates, which in Polypterus are 
placed below the inferior margin of the large one and behind the maxilla. The 
bone p. op. in Rhizodopsis may then be considered as the prwoperculum, the 
two others, x and x' , as equivalent to the cheek cuirass in Lepidosteus, or to 
the posterior set of sub-orbitals in other Lepidosteids (e.g., Lepidotus), and in 
the Palseoniscidse. 

In front of the bone x, and above the maxilla, there are in some specimens 
evident enough remains of the proper sub-orbitals, which seem to have cor- 
responded in number and position pretty closely to those in Osteolepis. Two 
of them (s.o. fig. 2)), corresponding respectively to the posterior-inferior and 
anterior-inferior parts of the boundary of the orbit, are clearly seen in many 
specimens, but the unfortunate manner in which the heads are crushed renders 
any further description hardly possible. 

The space between the right and left mandibular rami is occupied by a set 
of jugular plates. Professor Young 
has described these as consisting of 

"two pairs, principal and posterior," 

and has also stated that there is "no 

trace of median or lateral plates." * 

The specimens before me, however, do 

not corroborate the views above quoted. 

I find two principal jugulars (J. figs. 2 

and 3) occupying almost the whole of 

the space. Each of these is of the 

usual oblong shape, and broader behind 

than in front. The short and rounded 

posterior margin passes uninterruptedly 

into the internal one, which is more 

convex than the external for the 

greater part of its length ; near the 

front, however, the internal and ex- 
ternal margins converge and meet in an acute angle. What Professor Young 

means by a "posterior" jugular I am unable to determine, unless he 

has mistaken for such a plate the broad infra-clavicular element of the 

shoulder girdle, which, as in the recent Polypterus, is overlapped by the 

posterior margin of the principal jugular. The presence of lateral jugulars 

* Op. tit, p. 596. 



.op. 




:op. 



Fig. 3. —Under Surface of the Head of Rhizodopsis sauroides. 

mil. mandible ; ;'. principal jugular ; l.j. lateral jugular ; 

m.j. median jugular ; s.op. suboperculum. 



178 RAMSAY H. TRAQUAIR 

(/./.) is clearly shown in several specimens, and are at least five in number on 
each side. Of these, the hindermost is also the largest, and is situated below 
the lower margin of the suboperculum, extending also beyond the posterior 
margin of the principal jugular; the remaining four are placed between the 
last-named plate and the mandible, and diminish in size regularly from behind 
forwards. There is also the clearest possible evidence of a median jugular 
(m.j.), of a somewhat oval-acuminate form, placed immediately behind the 
symphysis of the mandible, and overlapping to some extent the anterior 
extremities of the principal jugulars. That the lateral and median jugular 
plates were not noticed by Professor Young, is clearly due to the more 
imperfect material then at his command. 

Conclusion. 

The foregoing investigation into the osteology of the head of Rhizodopsis, 
deficient as it is with regard to the more internally situated parts, nevertheless 
brings out, in a very striking manner, the affinity of that genus to the rhombic- 
scaled Saurodipterini, and supplies further evidence, were that now required, 
of the comparatively small value of the mere external forms of scales as 
indicating the natural affinities of ganoid fishes. 

No one acquainted with the structure of Megalichthys can fail to be struck 
with the extreme resemblance which its cranial osteology bears to that of 
Rhizodopsis, not only in general arrangement but in the shapes of individual 
bones, — a resemblance shared in as well by the teeth with their labyrinthically 
plicated bases, by the shoulder bones, by the fins in their structure and position, 
and by the vertebral column with its ring-shaped centra. Beyond a doubt, the 
affinities of Rhizodopsis are much more with the rhombiferous Saurodipterini 
than with the cycliferous Holoptychiiclse, although, on account of the form of 
the scales, both Rhizodopsis and Rhizodus were once included in the genus 
HoloptycMus. 

Very distinct family characters are, however, presented by the Saurodipte- 
rini in the scales having assumed a sharply rhombic contour, in their free 
surfaces, as well as those of the cranial bones and fin rays, being covered with 
a layer of brilliant ganoine, and in the tendency of many of the bones of the 
head to fusion with each other. In Megalichthys, for example, the mandible 
though closely resembling that of Rhizodopsis in external contour and in the 
form and arrangement of its teeth, has the elements — which in the latter genus I 
have designated as angular, dentary, infradentary , and internal dentary — all 
fused into one piece, an oblique line on the outside of the jaw usually indicating 
the original separation of the dentary. In some Old Red Sandstone Sauro- 
dipterini the original separation of the parietal, squamosal, and posterior frontal 



ON THE CRANIAL OSTEOLOGY OF RHIZODOPSIS. 179 

elements of the cranial buckler, is on the surface almost entirely obliterated. 
These circumstances would lead us to the conclusion that the Saurodipterini 
constitute a more specialised type than the Cyclodipteridse, in which, in a 
previous essay,"' I have included the genera Rhizodus, Rhizodopsis, Strepsodus, 
Archichthys, and Tristichopterus, the Glyptolaemini being probably intermediate. 

Note added July 20, 1881. — For the term " Cyclodipteridae," which I have 
hitherto used for the family to which Rhizodopsis, Rhizodus, &c, belong, and 
which I borrowed from Dr Lutken ( " Begrenzung und Eintheilung der 
Ganoiden," German edition, p. 47), though excluding the Holoptychii, which 
were also here placed by him, I propose in future to substitute " Rhizodon- 
tidae," as being in every way more appropriate. 

* "On the Structure and Affinities of Tristichopterus alatus," Trans. Roy. Soc. Ed., 1874. 



VOL. XXX. PART I. 2 E 



( 181 ) 



VIII. — On the Action of Phosphide of Sodium on Haloid Ethers and on the Salts 
of Tetrabenzyl-Phosphonium. By Professor Letts and N. Collie, Esq. 

The phosphines, or substances derived from phosphuretted hydrogen by 
the partial or complete replacement of its hydrogen by hydrocarbon radicals, 
have formed the subject of many valuable researches ; but although their 
discovery was anterior to that of the compound ammonias, their study has made 
comparatively little progress. This is no doubt mainly due to the difficulty 
attending their preparation, a fact which is immediately forced upon the notice 
of any one who wishes to investigate them. 

In spite of the undoubted analogies existing between phosphines and 
amines, the methods employed for obtaining the former are, with one excep- 
tion, different from those by which the latter are usually prepared. The 
reason for this we may find in the great differences between the elements 
phosphorus and nitrogen — differences which are in many cases still apparent 
in their compounds. Thus, phosphorus forms no compound with carbon 
analogous to cyanogen ; nor have any phosphorised bodies been obtained 
up to the present time analogous to the cyanides of hydrocarbon radicals. 
Neither has a phosphorised cyanic acid, (HCPO), nor its hydrocarbon salts 
been obtained. 

And we have another link wanting in the chain of analogies existing 
between nitrogen and phosphorus, in the absence of compounds of the latter 
element analogous to the nitro-bodies. Now, the amines are usually prepared 
by one or other of the four following processes : — 

1. Action of nascent hydrogen on the cyanide of a hydrocarbon radical. 

2. Action of caustic potash on the cyanate of a hydrocarbon radical. 

3. Action of nascent hydrogen on a nitro-body. 

4. Action of ammonia on a compound of a hydrocarbon radical with a 

halogen. 

For the reasons given above, the phosphines cannot be prepared by pro- 
cesses corresponding with the first three of these methods ; but Hofmann, 
in his masterly researches on these bodies, has shown that it is possible 
to directly replace hydrogen in phosphuretted hydrogen by hydrocarbon 
radicals, in a manner similar to that employed in the fourth of the above 
processes. 

But this is not the only process we possess for obtaining the phosphines, 

VOL. XXX. PART I. 2 F 



182 PROFESSOR LETTS AND N. COLLIE ON THE 

although it is the only one analogous to any of those employed for preparing 
amines ; and we shall give a short sketch of the other methods by which, from 
time to time, the phosphines have been prepared. 

Paul Thenard was the discoverer of the first organic phosphorus com- 
pounds.""" In the year 1843 he investigated the action of chloride of methyl on 
phosphide of calcium ; and in 1847 he communicated to the Academy further 
results as to the nature of the bodies obtained in the reaction. The investiga- 
tion was attended with great difficulties, owing to the labour involved in 
separating the different products, and in obtaining them in the pure state ; also, 
on account of their explosive and inflammable nature, and their poisonous 
properties. 

In spite, however, of these difficulties, Thenard appears to have isolated 
trimethyl-phosphine ; a substance analogous to kakodyle, P 2 (CH 3 ) 4 ; and 
a substance analogous to solid phosphide of hydrogen, P 4 (CH 3 ) 2 . The 
last he describes as an inert solid body ; but the second, as a spontane- 
ously inflammable liquid boiling at 250° C. — very explosive, poisonous, and 
unstable. 

Thenard recognised the relations existing between trimethyl-phosphine 
and ammonia, and predicted the existence of the then undiscovered organic 
compounds of nitrogen, arsenic, and antimony. In the meantime, Wurtz and 
Hofmann had verified Thenard's predictions, having discovered the compound 
ammonias ; and Loewig and Schweitzer had obtained stib-ethyl. 

Hofmann and CAHOURSt turned their attention in 1855 to the study of the 
phosphines, and repeated Thenard's experiments, with this difference, however, 
that they substituted phosphide of sodium for phosphide of calcium. They 
obtained trimethyl-phosphine, Thenard's phosphorus kakodyle and iodide of 
tetramethyl-phosphonium — but only after great difficulty. Speaking of the action 
of phosphide of sodium on iodide of methyl, they say, — " The action is very 
energetic when the two are heated together (a chand). Moreover, inflammable 
and detonating substances are formed, so that this method of preparation is 

not without danger, and exposes the fruit of one's labour to loss It 

is unreliable {irop pen stir), and furnishes mixtures, of which the separation 
presents enormous difficulties." For these reasons, they sought for a simpler 
and more certain process. This they found in the action of zinc ethers 
on terchloride of phosphorus, which gives a compound of chloride of zinc 
and the tertiary phosphine, from which potash separates the latter in a state 
of purity. 

By means of this reaction, Hofmann and Cahours prepared the tertiary 
phosphines of methyl, ethyl, and amyl. They determined some of their most 

* Comptes Kendus, vols. xxi. and xxv. 
f Comptes Rondus, xli. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 183 

important properties, and showed that in many respects they resemble the 
tertiary amines, especially in the readiness with which they combine with 
iodides of hydrocarbon radicals to give quaternary compounds. They found, 
however, that, unlike amines, tertiary phosphines are capable of directly 
combining with oxygen. 

Hofmann continued the study of the tertiary phosphines, and communi- 
cated the results of his experiments to the Eoyal Society* in 1860. He 
confined his experiments chiefly to triethyl-phosphine, and, in his lengthy 
memoir, describes accurately its properties and reactions. He prepared 
and analysed oxide of triethyl-phosphine and the characteristic red com- 
pound which bisulphide of carbon forms with the phosphine itself, and he 
investigated the action of the latter on a considerable number of organic 
compounds. 

BERLEt attempted to obtain triethyl-phosphine by the action of phosphide 
of sodium on iodide of ethyl. The phosphide of sodium he prepared by the 
heating sodium and phosphorus together in rock oil. Iodide of ethyl only 
acted upon this at a high temperature, and he obtained only very small quantities 
of the tertiary phosphine. BerleJ next attempted to prepare the tertiary 
phosphine by heating sodium, phosphorus, and iodide of ethyl together in a 
sealed tube ; but although the bodies reacted, he does not seem to have 
obtained any very satisfactory results. 

Cahours, in 1859, prepared iodide of tetrethyl-phosphonium by the action of 
iodide of ethyl, on crystallised phosphide of zinc (prepared by heating the 
metal in phosphorus vapour) at 180° C. The next experiments on the pre- 
paration of phosphines are very interesting and important. 

Previous to these only tertiary and quaternary compounds had been 
obtained, but Hofmann § showed in an elegant manner that the primary and 
secondary bases may be formed by the action of phosphuretted hydrogen on 
the iodides of hydrocarbon radicals — a process exactly analogous to that 
employed by him for preparing the corresponding amines. Phosphuretted 
hydrogen, however, does not behave in exactly the same manner as ammonia 
in this reaction, for Hofmann found that the replacement of hydrogen does 
not proceed further than the second atom; whereas with ammonia all the 
hydrogen is replaced step by step, and even quaternary compounds are 
formed. 

Moreover, ammonia acts on the iodides of hydrocarbon radicals much more 
readily than phosphuretted hydrogen, and at lower temperatures. 

* Transactions Eoyal Society, London, vol. cl. p. 409. 

t Journ. fur. prac. Chem., lxvi. p. 73. 

X Comptes Eendus, xlix. 

§ Berichte der. deutsch. chem. Ges., iv. pp. 205, 372; v. p. 100. 



184 PROFESSOR LETTS AND N. COLLIE ON THE 

Hofm ann's process for obtaining primary and secondary phosphines — 
which he employed successfully in the methyl, ethyl, and benzyl series 
— consists in heating a mixture of phosphonium iodide, zinc white, and the 
hydrocarbon iodide, in sealed tubes for some hours at a temperature of 
160°-180°. The tubes are then found to contain a white crystalline mass, 
consisting of compounds of the hydriodates of the primary and secondary bases 
with zinc iodide. 

The reactions which occur are represented by the equations, 

2C 2 H 5 I + 2PH 4 I + ZnO = 2[(C 2 H 5 )H 2 P,HI], Znl 2 + H 2 . 
2C 2 H 5 I + PH 4 I + ZnO = (C 2 H 5 ) 2 HP, HI, Znl 2 + H 2 . 

The separation of the primary from the secondary compound is accom- 
plished with the greatest ease. It is only necessary to add water to the con 
tents of the sealed tubes when the compound of the primary base is decom- 
posed and the base itself set at liberty. When it has been distilled off, the 
addition of potash to the residue separates the secondary base. 

Hofmann also studied the action of phosphuretted hydrogen on the alcohols 
at a high temperature, and with a singularly interesting result. 

Not only does phosphuretted hydrogen act on the alcohol, but the bodies 
produced consist entirely of tertiary and quaternary compounds, no primary 
or secondary compounds being formed at all. Thus the action of phos- 
phuretted hydrogen on an iodide of a hydrocarbon radical is exactly comple- 
mentary to its action on an alcohol. 

In employing the action of phosphuretted hydrogen on ordinary alcohol 
for the preparation of the tertiary and quaternary phosphines, Hofmann places 
iodide of phosphonium at the bottom of a sealed tube, and above it the 
alcohol in a smaller tube. The vapour of the phosphonium iodide thus comes 
in contact gradually with the alcohol. The reaction is complete after six to 
eight hours digestion at 180°. The tubes are then found to be full of a white 
crystalline mass, from which caustic potash liberates the tertiary phosphine, 
whilst the iodide of the phosphonium remains in solution. 

The reactions which occur are represented by the equations 

3(C 2 H 5 OH) + PH 4 I = P(C 2 H 5 ) 3 HI + 3H 2 . 
4(C 2 H 5 OH) + PH 4 I = P(C 2 H 5 ) 4 I + 4H 2 . 

Michaelis* has comparatively recently added to our knowledge of the 
phosphines, and to the methods of preparing them. 

By passing the mixed vapours of terchloride of phosphorus and benzol 
through a red hot tube he obtained j^osphenyl-chloride, 

PC1 3 + C 6 H 6 = (C 6 H 5 )PC1 2 + HC1 . 

* Liebi^s Annalen, 181, p. 280. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHEES. 185 

a substance which he also prepared by the action of terchloride of phos- 
phorus on mercury cli-phenyl, 

PC1 3 + Hg(C 6 H 5 ) 2 = (C 6 H 5 )PC1 2 + HgCl(C 6 H 5 ) . 

By the action of water on this body, phosphenylous acid is produced, 

(C 6 H 5 )PC1 2 + 2H 2 = (C 6 H 5 )P0 2 H 2 + 2HC1 , 

and this when destructively distilled yields phenyl-phosphine — the phosphorus 
analogue of aniline, 

3 {(C 6 H) 5 P0 2 H 2 f = (C 6 H 5 )PH 2 + 2C 6 H 6 + 2HP0 3 . 

The same body results when hydriodate of phosphenyl-iodide (obtained by 
the action of hydriodic acid on phosphenyl-chloricle) is decomposed with 
alcohol. 

We were led in the first instance to the experiments to be presently 
described by the difficulty which one of us had experienced in preparing 
triethyl-phosphine on the large scale. Hofmann's later method had been at 
first resorted to, but in spite of numerous experiments, it had led to no satis- 
factory results. The pressure produced when alcohol and iodide of phos- 
phonium are heated together is enormous, especially at the high temperature 
(180° C.) at which they react, and in almost nine cases out of ten it was found 
that the sealed tubes burst. 

Nor is the other process for preparing triethyl-phosphine, viz., by treating 
zinc-ethyl with terchloride of phosphorus, a simple operation. The preparation 
of zinc-ethyl is expensive and troublesome, and although it reacts readily with 
the terchloride, the reaction is not so simple as might be expected. Scarcely 
50 per cent, of the theoretical quantity of crude phosphine can be obtained, 
and this crude product contains impurities in considerable quantities, which 
are very difficult to remove. The preparation of triethyl-phosphine is in fact 
an expensive, uncertain, and troublesome operation. 

Such being the case, and one of us requiring large quantities of it, the 
question naturally presented itself— Is there no simpler and less expensive 
process for preparing a tertiary phosphine % It seemed to us that one of the 
processes — and in fact the earliest — for preparing these bodies ought to be an 
extremely good one, if the difficulties attending its general application could be 
removed. The process to which we allude depends upon the ease with which 
metallic phosphides can be formed, and the readiness with which haloid ethers 
act on them. As before stated, Thenard, Berle, Cahours, Hofmann and 
others, have worked with this process, but it has not met with great favour, 
and was abandoned by Hofmann (who employed phosphide of sodium) on 



186 PROFESSOR LETTS AND N. COLLIE ON THE 

account of the uncertainty of the reaction, the frequent explosions, and the 
great difficulties in separating the resulting phosphines from each other, — " not 
to speak of the difficulty of obtaining the phosphide of sodium fit for the 
reaction." 

It seemed to us, however, that in phosphide of sodium an admirable reagent 
was at hand for the preparation of tertiary phosphines — provided only, to quote 
again Hofmann's words, that it can be obtained in a state "fit for the 
reaction." 

This conclusion has been borne out by our experiments. With proper 
precaution, phosphide of sodium may be obtained in any quantity, and in a 
perfectly safe condition. It reacts with haloid ethers in a perfectly smooth 
manner, nor have we ever had an explosion, nor remarked the production of 
explosive bodies. 

Our first experiments were made with iodide of ethyl. The reaction occurs 
at ordinary temperatures with ease, the iodide of ethyl boils violently, and the 
chief product of the reaction appears to be the iodide of tetrethyl-phosphonium. 
We have not as yet, however, brought these experiments to a conclusion, 
because of the difficulties which we experienced in separating the phosphines 
and phosphonium salt from the iodide of sodium produced along with them in 
the reaction. 

Our next experiments were made in the benzyl series, which we chose 
partly because neither tribenzyl-phosphine nor tetrabenzyl-phosphonium salts 
have hitherto been obtained, and partly because no deliquescent or volatile 
bodies were likely to be formed, thus rendering the investigation free from 
those difficulties which cause experiments in the methyl and ethyl series to be 
so troublesome and laborious. To these reasons for our choice of benzyl must 
be added its similarity to fatty radicals and the well-known ease with which its 
compounds react. 

Before proceeding to describe our experiments on the preparation of phos- 
phide of sodium, and on its action on chloride of benzyl, we consider it 
necessary to give a short account of Hofmann's researches on monobenzyl- 
and dibenzyl-phosphine, which we believe to be the only ones that have been 
made on benzyl-phosphines. 

Benzyl- Phosphines. 

The following is an abstract of Hofmann's paper on "Aromatic Phos- 
phines":"" — 

He was induced to experiment on the aromatic series, in consequence of the 
readiness with which, by the use of iodide of phosphonium, he had obtained 

* Hofmann, Ber. d. deutsch. chem. Ges., iv. p. 100. 



ACTION OF PHOSPHIDE OF SODIUM ON" HALOID ETHERS. 187 

methyl- and ethyl-phosphines. His first attempts were made with the view of 
obtaining a phenyl-phosphine analogous to aniline, a substance highly interest- 
ing from a theoretical point of view. 

To obtain this body he heated, under varying conditions, phenyl-chloricle 
and iodide of phosphonium ; but the experiments did not lead to a successful 
result, the phenyl-chloride becoming reduced to benzol, which even at high 
temperatures was not further acted upon. That the reaction did not proceed in 
the desired manner was, as he says, not surprising, considering the inertness of 
chloride of phenyl and the fact that aniline cannot be obtained by acting 
on it with ammonia. 

Equally unsuccessful were his efforts to obtain the tertiary phosphine and 
the quaternary compound by the action of phenol upon iodide of phosphonium, 
though phosphorised bodies resulted, the nature of which he did not ascertain. 

Experiments to obtain a phosphorised toluidine led to no successful 
issue ; but, on the other hand, the preparation of a phosphorus analogue of 
benzylamine presented no difficulty, as indeed he did not doubt, considering 
the readiness with which chloride of benzyl reacts with ammonia. 

Benzyl Phosphine, C 7 H 7 PH 2 . — This body is formed when chloride of 
benzyl (which may be employed in the crude condition) is heated for six hours 
at a temperature of 160° with a mixture of phosphonium iodide and zinc 
oxide. The substances are taken in the proportions of 2 molecules benzyl 
chloride, 2 of phosphonium iodide, and 1 of zinc oxide. 

When complete reaction has occurred the sealed tubes in which the mixture 
has been heated contain a white crystalline mass. On opening them a large 
quantity of phosphuretted hydrogen is evolved. On distilling the product of 
the reaction with water a heavy, oily liquid passes over of highly characteristic 
odour. This is separated, dried with caustic potash, and distilled in hydrogen. 
The thermometer rises to 1 80°, and then remains stationary, whilst a considerable 
quantity of a colourless, highly refractive liquid distils. This is monobenzyl- 
phosphine, whilst the lower boiling fraction consists mainly of toluol, and the 
residue in the retort contains dibenzyl-phosphine and other products. A 
simple distillation of the crude benzyl-phosphine thus obtained gives the pure 
body boiling at 180° C. In its properties it resembles in the main other 
primary phosphines. It oxidises in contact with the air, its temperature 
rising to 100° C, thick white vapours being formed. It is insoluble in water, 
but easily soluble in alcohol and ether. It forms a colourless crystalline 
hydriodate, only slightly soluble in fuming hydriodic acid, and which, like other 
salts of primary phosphines, decomposes in contact with water into hydriodic 
acid and the free base. Monobenzyl-phosphine also combines with hydro- 
chloric acid, and gives a yellow insoluble chloro-platinate. 



188 PROFESSOR LETTS AND K". COLLIE ON THE 

Benzyl-phosphine is formed, according to the reaction, 

2C 7 H 7 C1 + 2(H 3 PHI) + ZnO = 2[(C 7 H 7 )H 2 PHI], ZnCl 2 + H 2 0. 

This equation, however, only expresses one phase of the reaction, in which 
simultaneously with monobenzyl-phosphine other products are formed. Of 
these Hofmann could only succeed in isolating — 

Dibenzyl-Phosphine. — This compound is contained in the residues after 
the benzyl-phosphine has been distilled off. On long standing in contact with 
solid potash, these solidify to a soft crystalline mass, which is collected on a 
linen filter, well squeezed, dissolved in alcohol, and the solution decolorised 
with animal charcoal. It then deposits, on cooling, beautiful white crystals 
of the dibenzyl-phosphine, which only require to be recrystallised from alcohol 
to obtain them in a state of purity. 

Dibenzyl-phosphine thus obtained forms glistening white needles of large 
size, which are tasteless and colourless. They are insoluble in water, spar- 
ingly soluble in cold alcohol, readily in hot alcohol. The crystals melt 
at 205°, and sublime at higher temperatures, but not without partial decom- 
position. 

With the introduction of the second benzyl group the alkaline characters, 
which are perfectly apparent in monobenzyl-phosphine, disappear. 

Dibenzyl-phosphine dissolves in no acid, nor could Hofmann obtain its 
chloro-platinate. Its composition could, therefore, only be determined by an 
analysis of the purified substance. The formation of dibenzyl-phosphine may 
be expressed by the equation, 

2C 7 H 7 C1 + H 3 PHI + ZnO = [(C 7 H 7 ) 2 HP,HI], ZnCl 2 + H 2 . 

The mother liquors of the dibenzyl-phosphine contain another phosphorised 
body, which Hofmann suspected to be tribenzyl-phosphine, but in spite of 
many efforts he was not able to prepare that body. 



ACTION OF PHOSPHIDE OP SODIUM ON" HALOID ETHERS. 189 

Phosphide of Sodium. 

The success attending the use of this reagent for the preparation of phos- 
phines depends entirely on the observance of certain conditions in its manu- 
facture, which we have carefully determined by experiment. 

The most essential of these are, the manner in which the sodium and phos- 
phorus are allowed to combine, and the nature and quantity of the reagent em- 
ployed to prevent the temperature from rising too high during their combination, 
and to protect the resulting phosphide from subsequent oxidation by the air. 

If proper attention be paid to these precautions, phosphide of sodium may 
be prepared in a perfectly safe manner in large quantities and of uniform com- 
position. Whatever this latter may be, the bulk of the phosphide behaves as 
Na 3 P, as is evident from the nature of the products obtained from it. 

We prepare the phosphide by melting sodium under xylol and adding 
ordinary phosphorus in small pieces, — shaking the vessel in which the mixture is 
made after each addition of the latter substance in order to bring it thoroughly 
into contact with the sodium. As regards the proportions of sodium and phos- 
phorus, the theoretical quantities for the formation of the compound Na 3 P, are, 
in round numbers, 7 of sodium and 3 of phosphorus ; but our experiments show 
that with these proportions very bad results are obtained, as regards phosphine 
compounds. It appears to be necessary to employ a large excess of phosphorus 
— the proportions which we have found to be most suitable being 19 of that 
body to 20 of sodium — very nearly twice the theoretical quantity. 

Considering that large quantities of phosphorus are converted into the 
amorphous modification, which does not apparently combine with sodium at 
those temperatures at which we operate, a simple explanation is afforded of the 
necessity for employing an excess of phosphorus ; for, otherwise, unchanged 
sodium remains in the product, and, as we shall presently explain, this readily 
decomposes the chief phosphine compound produced when phosphide of sodium 
acts on chloride of benzyl. But, on the other hand, an excess of phosphorus 
over the proportions we have indicated also exercises a prejudicial influence 
on the quantity of phosphine compounds produced, and at present we can only 
explain this on the assumption that higher phosphides of sodium are formed, 
which are inactive. 

The xylol necessary to prevent the reaction of the phosphorus and sodium 
from being too violent must be free from water and high boiling impurities. A 
single distillation of ordinary xylol is sufficient to purify it for this purpose. 
The proportion of xylol employed is also an important condition. This, how- 
ever, we shall best consider when we describe our experiments on the action of 
the phosphide on chloride of benzyl. 

The following is a description of the exact method we employ for preparing a 
batch of phosphide of sodium containing 78 grms. of sodium and phosphorus : — 

VOL. XXX. PART I. 2 G 



190 PROFESSOR LETTS AND N. COLLIE ON THE 

40 grins, of freshly scraped sodium are placed in a flask of 500 cc. capacity,— fitted 
with a cork and wide glass tube placed vertically, and about 4 feet in height, to 
serve as a reversed condenser — and to it 200 grins, of xylol are added. The 
mixture is then warmed on a sand-bath until the sodium has melted. 38 grms. 
of phosphorus are cut into pieces the size of a pea, and these are placed in an 
evaporating basin containing xylol. They are then added gradually to the 
melted sodium. The first addition of the phosphorus causes a very violent 
reaction, so that care is necessary to prevent the xylol from boiling over. As 
soon as the xylol boils, the cork with its tube to serve as condenser is fitted into 
the flask, and the reaction allowed to proceed until the boiling ceases. More 
phosphorus is then added, and the flask well shaken after each addition. It is 
not necessary to heat the mixture, as the temperature remains sufficiently high 
to keep the reaction in progress to the end. The resulting product is a granular 
black powder, which is completely prevented from change by the xylol which 
surrounds it. We have kept a loosely corked flask containing it for upwards 
of six months without noticing that it had suffered change. But the slightest 
trace of moisture at once acts upon it, liberating spontaneously inflammable 
phosphuretted hydrogen. 

Action of Chloride of Benzyl on Phosphide of Sodium. 

Chloride of benzyl when boiled with the phosphide of sodium — prepared in 
the manner we have described — readily acts on it with the formation of common 
salt and compounds of benzyl and phosphorus. The main phosphine product 
is the chloride of tetrabenzyl-phosphonium, which may be obtained in large 
quantities by this method, and is easily isolated from the products of the re- 
action. Other phosphorised bodies are, however, formed in considerable 
quantity, the nature of which we shall discuss later. 

As we have before stated, careful attention must be paid to the preparation 
of the phosphide of sodium, otherwise, as we have often observed, no chloride 
of the phosphonium results, although the whole of the phosphide of sodium may 
be acted on. On the other hand, the phosphide sometimes remains unacted on, 
or the reaction takes five or six times longer than is necessar}^ in a properly 
conducted experiment. A very important condition of success is the right pro- 
portion of xylol employed in the preparation of the phosphide of sodium. 
We have found it advisable to take an excess of this substance in the first 
place, but to distil off this excess before adding the benzyl chloride. 

With the quantities of sodium, phosphorus, and xylol mentioned already 
we distil off (in an oil bath) 170 grms., that is to say, a sufficient quantity 
to leave the phosphide of sodium almost dry, but in such a condition that 
the addition of a small quantity of benzyl- chloride may thoroughly wet 
the mass and permit of its being boiled without becoming superheated at 
the bottom. The chloride of benzyl is placed in a tap funnel, and allowed to 









ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 



191 



drop clown the wide tube used as a reversed condenser in the preparation of 
the phosphide of sodium. The tap is turned so that the benzyl-chloride may- 
drip fast on to the phosphide of sodium contained in the flask (in which it has 
been made), which is heated in an oil bath. Care must be taken, however, 
lest the benzyl-chloride run in too rapidly, as in that case the ebullition (caused 
by the reaction) may be so violent that liquid collects in the condensing tube 
and is blown out by the vapour. 

That reaction is occurring in the proper manner may be known by the 
change of colour of the phosphide of sodium from black to orange. It is 
advisable after the reaction has proceeded for some time, to remove the flask 
from the oil bath and to shake it vigorously ; but as the lower portions of the 
phosphide are apt to become superheated, the shaking must be done cautiously. 



PREPARATION OF TETRABENZYL-PHOSPHONIUM CHLORIDE. 



Benzyl-Chloride.. 

Xylol* 

Sodium 

Phosphorus 



Unacted on henzyl- ) 
chloride and xylol... J 

Yield of (C 7 H 7 ) 4 PClt.... 

Theoretical yield , 



1 


2 


3 


4 


5 


6 


Grms. 


Grms. 


Grms. 


Grms. 


Grms. 


Grms. 


100 


160 


200 


115 


220 


220 


100 


160 


Toluol 


Toluol 


500 


500 


14 


22-5 


28 


17 


31 


31 


6 


21 


26 


15 


28 


28 


12 


23 


7 


7 


12 


88 


141 


176 


101 


193 


193 



Grms. 
100 

100-40 

14 



Grms. 
100 

100-60 

14 



Flask broke 4 



9 


10 


11 


Grms. 


Grms. 


Grms. 


100 


200 


100 


100-55 


100-75 


100-55 


18 


28 


21 


17 


26 


19 


23 


38 


17 


88 


176 


88 



12 



Grms. 
100 

100-60 

20 

19 



38 



Benzyl-Chloride . 

Xylol* 

Sodium 

Phosphorus 



Unacted on benzyl- \ 
chloride and xylol j 

Yield of (C 7 H 7 ) 4 PC1 +.. 

Theoretical yield .... 



13 



Grms. 
100 

100-25 

20 

19 

80 

18 



14 


15 


16 


17 


18 


19 


Grms. 


Grms. 


Grms. 


Grms. 


Grms. 


Grms. 


100 


100 


200 


100 


200 


200 


100-25 


100-25 


245-200 


120-100 


200-170 


200-180!! 


20 


20 


35 


20 


40 


40 


19 


19 


33 


19 


38 


38 


90 


90 


78 


35 


40 


60 


7 


5 


20 


43 


70 


70 


88 


88 


176 


88 


176 


176 



20 



Grms. 
200 

230-205 

40 

38 

30 

40 
176 



21 



Grms. 
200 

220-175 

40 

38 

52 

65 
176 



22 



Grms. 
200 

220-178 

40 

38 

58 

137 
176 



23 



Grms. 
200 

220-176 

40 

38 

75 

137 
176 



* In 7 and subsequent experiments a quantity of the xylol was distilled off after the preparation of the phosphide of 
sodium, and before the chloride of benzyl was allowed to act on it. The numbers after the sign - show how much of the 
xylol was thus distilled off. 

+ The chloride was squeezed in a cloth filter until it was as dry as possible and then weighed. 



192 PROFESSOR LETTS AND N. COLLIE ON THE 

When the reaction is at an end only a few black particles are visible— the 
solid portions of the product appearing of a dirty orange colour. 

The preceding table shows the results of our experiments on the influence 
of the conditions which affect the yield of chloride of tetrabenzyl-phosphonium, 
which no doubt will be serviceable in experiments on the phosphines of other 
radicals. 

Extraction of Chloride of Tetrabenzyl-Phosphonium from the Product. 

The product of the reaction we have just described consists of a mixture of 
common salt, a small quantity of phosphide of sodium which remains unacted 
on, xylol, chloride of benzyl, and phosphorised bodies, of which the chief is the 
chloride of tetrabenzyl-phosphonium. 

To extract the latter we have tried several different processes, the simplest 
and best of which is the following : — 

The flask containing the product of the reaction is connected with a con- 
denser, and heated in an oil bath at a temperature of 180°-200° C* as long as 
liquid distils. The mass remaining in the flask is then detached by a glass 
rod, and added by small portions at a time to cold water acidulated with 
hydrochloric acid ; but as spontaneously inflammable phosphuretted hydrogen is 
evolved, the operation must be conducted cautiously. When this gas ceases to 
come off, the solution is filtered and the residue boiled with a considerable 
quantity of water (2 or 3 litres). The aqueous extract is then filtered (through 
a cloth filter), more water is added to the residue, and the two boiled together 
as before, the solution filtered, and this treatment repeated until a portion of 
the extract gives no crystalline precipitate when cooled and mixed with hydro- 
chloric acid. The united aqueous extracts are then mixed with about 10 per 
cent, of their volume of strong hydrochloric acid and allowed to cool. Almost 
every trace of chloride of tetrabenzyl-phosphonium is then precipitated in the 
form of minute needles. The compound is collected on a linen filter, well 
squeezed to free it from mother liquor, and recrystallised from boiling water, 
from which it separates on cooling in beautiful needles often an inch and a half 
long, and almost perfectly white and pure. The residue which remains after ex- 
tracting the phosphonium compound forms a reddish-brown insoluble solid mass. 
It contains phosphorised bodies, the nature of which we shall consider later. 

Chloride of Tetrabenzyl-Phosphonium. — As this body forms the starting 
point from which the other salts of the phosphonium are obtained, we have 
examined its properties very carefully. It is sparingly soluble in cold water. 
(In a rough experiment 100 cc. of water at ordinary temperatures dissolved 
0*35 grm.) It is much more soluble in boiling water, and crystallises from 

* A higher temperature must be avoided, as we find that it materially diminishes the yield of phos- 
phonium salt. 



ACTION OF PHOSPHIDE OB' SODIUM ON HALOID ETHERS. 183 

it on cooling in long needles, which are somewhat thick and are perfectly 
colourless. The addition of common salt, and especially of hydrochloric acid, 
to its aqueous solution reduces its solubility in a remarkable manner : in fact, 
it is practically insoluble in the dilute acid, and traces may be detected in 
solution by this means. Chloride of tetrabenzyl-phosphonium crystallises from 
an aqueous solution with two molecules of water of crystallisation. The dried 
salt is readily soluble in alcohol, and the solution on slow evaporation yields 
very beautiful, colourless, rhombic crystals of considerable size and perfect 
symmetry. These consist of the anhydrous compound. The dried salt is also 
easily dissolved by chloroform, and on evaporation, the solution yields large 
colourless crystals, which lose chloroform on exposure to the air, and appear to 
have effloresced. A determination of the loss which a specimen of the salt 
crystallised from chloroform suffered when heated to 100° C. showed that the 
chloride had crystallised with 1 molecule of chloroform. We are not aware 
that " chloroform of crystallisation " has been noticed before in any substance. 

Chloride of tetrabenzyl-phosphonium fuses at 224° C, and volatilises slowly 
at that temperature apparently unchanged. Heated more strongly it decom- 
poses, yielding solid and liquid products, amongst the former being free 
phosphorus. Boiled with a solution of potash it suffers a change, which we 
shall refer to later (see p. 204). 

The composition of the chloride was determined by estimations of carbon, 
hydrogen, chlorine, phosphorus, and water. The combustion was made with 
chromate of lead and oxide of copper. The chlorine was determined gravi- 
metrically by j)recipitation as chloride of silver, and the phosphorus by titration 
with uranium solution after the salt had been fused with a mixture of caustic 
potash and nitrate of potassium. 

Analysis of Chloride, of Tctrabenzyl-Phosphonium. 

Dried Salt. 

294 grm. gave "837 carbonic anhydride = 77*7 per cent, carbon. 

294 „ „ -188 water 
•6943 „ „ -2298 chloric 
•6625 „ „ by titration 



Carbon, 
Hydrogen, . 
Chlorine, 
Phosphorus, 



* The hydrogen in this experiment is too high and the carbon too low, probably because the salt 
had not been completely dried. 



'{■ 



of silver = 


7T 
8-2 


„ hydrogen. 
,, chlorine. 


• = 


6-6 


„ phosphorus. 


Obtained, 

77-7 




Calculated for 
(C 7 H r ) 4 PCl. 

78-0 


7-1 


. 


6-5 


8-2 




8-2 


6-6 


• 


7-2 


99-6 




99-9 



194 PROFESSOR LETTS AND N. COLLIE ON THE 

Crystallised Salt. 

(1) 26-055 grras. lost at 110° 2-03 grms. water = 77 per cent. 
2)* 3-2392 „ „ -258 = 8-0 

Obtained. .-, , , , , ,. 

Calculated tor 

i. ii. (C 7 H 7 ) 4 PC1,2H 2 0. 

Water, 77 8-0 77 

Crystallised from Chloroform. 

4'2015 grms. lost at 110° -8938 grm. chloroform = 21 - 2 per cent. 

Calculated for 
Obtained. (C 7 H 7 ) 4 PC1,CHC1 3 . 

Chloroform, 21-2 217 

Chloroplatinate of Tetrdbenzyl-Phosphonium. — This compound is easily 
formed on mixing aqueous or alcoholic solutions of chloride of platinum and 
chloride of tetrabenzyl-phosphonium. In the former case a light orange 
precipitate is produced, which appears to be amorphous to the naked eye. In 
the latter, if the two solutions are boiling and dilute, a distinctly crystalline 
precipitate is produced of a darker colour. 

The chloroplatinate is very insoluble, almost absolutely so in water, and 
is only very slightly soluble in alcohol. 

The composition of the salt was determined by a combustion with chromate 
of lead and oxide of copper — as in all chloroplatinates of phosphines and 
phosphoniums the determination of chlorine and platinum is very tedious and 
uncertain. 

If attempts are made to determine platinum by the weight of the residue 
left on ignition, the results are often as much as 2 per cent, too high, owing to 
the formation of phosphide of platinum. Precipitation by sulphuretted hydrogen 
we have found to be always incomplete with chloroplatinates of benzyl- 
phosphines 

Hofmann's methodt gave us results which were too low. We could, in fact, 
discover no process for estimating the platinum in this salt, or the chlorine, 
and were in consequence obliged to fix its composition by an organic analysis. 
The results of this are as follows :— 

•348 grm. gave 711 of carbonic anhydride = 557 per cent, carbon. 
•348 „ „ # 1508 of water = 4 - 8 per cent, hydrogen. 

Obtained. Calculated for {(C 7 H 7 ) 4 PCl} 2 PtCl 4 . 

Carbon . . . 557 .... 56-0 
Hydrogen ... 4*8 .... 4*7 

* In this experiment the salt was heated for a whole day, and probably some of it volatilised, 
f Trans. Roy. Soc. Lond., vol. cxlvii. p. 575. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 195 

Sulphate of Tetrabenzyl-Phosphonium. — Nine grms. of the phosphonium 
chloride were dissolved in about a pint of hot water, the solution was 
then cooled somewhat, and sulphate of silver added until it ceased to be 
converted into chloride. The solution was then filtered and concentrated, 
and yielded a crop of very beautiful colourless rhombic plates of considerable 
size.* A determination of water of crystallisation and of sulphuric acid 
showed that these consisted of the normal sulphate crystallising with six mole 
cules of water. 

Salt dried at 120°. 

Sulphuric Acid (1 and 2 volumetrically by the chromate method, 3 gravimetrically). 

(1) 0-5745 grm. required 2-6 cc. J normal BaCl 2 =0-0520 SO 3 =9-0 per cent. 

(2) 0-4203 „ „ 3-8 „ £ „ „ =0-038 „ =9-0 „ 
(3)0-5053 „ gave 0131 grm. BaS0 4 =0-045 „ =8-9 „ 

Air-dried Salt. 

Sulphuric Acid. 

0-6865 grm. required 5'6 cc. £ normal BaCl=0-056 SOj=8 - 2 per cent. 
Water (by heating to 125° C). 

0-5704 lost 0-0574 H 2 O = 10-0 per cent. 

Obtained. 



Dry Salt. Undried. 

I. II. in. Salt. 

Sulphuric anhydride, . 9'0 9-0 8-9 8-2 

Water, .... — — — 10-0 

Calculated. 

{(C 7 H 7 ) 4 P} 2 S07~'T(C 7 H 7 ) 4 P} 2 S0r6H 2 . 
Sulphuric anhydride, . . 9'0 8 - l 

Water, .... — 10-8 

The normal sulphate is one of the most soluble salts of the phosphonium. 
It has no definite melting point. It grows pasty at 195°, but is not completely 
fused until 220°. 

Action of Sulphuric Acid on Chloride of Tetrabenzyl-Phosphonium. — When 

* In our notice of these experiments in the " Proceedings," we stated that the action of sulphate of 
silver on the phosphonium chloride yields the acid sulphate, and not the normal sulphate of the 
phosphonium. There can be but little doubt that we did obtain and analyse the acid salt, but its forma- 
tion was probably due to the fact that we acted on a boiling solution of the chloride with sulphate of 
silver, and the filtered solution therefore contained a considerable quantity of the latter salt which we 
decomposed by sulphuretted hydrogen, thus setting free sulphuric acid which combined with the normal 
sulphate ; and as the acid sulphate thus formed is much less soluble than the normal sulphate, it is 
probable, as Ave analysed the first crop of crystals, that they consisted entirely of the acid sulphate. 
We have verified the correctness of this explanation by an experiment conducted in a similar manner. 



196 PROFESSOR LETTS AND N. COLLTE ON THE 

oil of vitriol is poured on to the chloride, an immediate effervescence occurs 
and clouds of hydrochloric acid are produced. Heated on the water bath, the 
mixture grows dark and eventually becomes completely liquid, provided that 
an excess of sulphuric acid has been taken. As soon as the whole of the 
hydrochloric acid has been driven off, the mixture is poured into a small 
quantity of cold water. A solid substance of a grey colour is precipitated, 
which by washing with a little cold water becomes free from sulphuric acid. 
It is then boiled with a tolerably large quantity of water, when it almost 
completely dissolves, leaving, however, a slight residue containing most of 
the black colouring matter. On cooling, almost colourless crystals are 
deposited, which are apparently rhombohedral plates united into spear-shaped 
forms. 

A single recrystallisation gives the new compound in a state of purity. 

Its analysis shows it to be the — 

Acid Sulphate of Tetralenzyl-Phosphonium. — The chloride has therefore 
been acted on by sulphuric acid in the same manner as common salt. Its 
analysis gave the following numbers : — • 

•397 grin, gave -9907 C0 2 — - 681 per cent, carbon. 

'•397 „ „ -2176 H 2 = 61 „ hydrogen. 

•5252 „ „ -2445BaS0 4 = 16-0 „ sulphuric anhydride. 

Obtained. Calculated for (C 7 H 7 ) 4 PHS0 4 . 

Carbon, . . . 681 . . . 68'3 

Hydrogen, ... 61 . . . 5 - 9 

Sulphuric anhydride, . 16"0 . . . 16'2 

Acid sulphate of tetrabenzyl-phosphonium is rather more soluble in water 
than the chloride, and very much less so than the normal sulphate ; it readily 
crystallises from hot water. The addition of hydrochloric acid to its solution 
causes the precipitation of the chloride. It has a sour taste, and its solution 
reddens litmus paper. 

It does not suffer any loss when heated to 100° C, and therefore contains 
no water of crystallisation. 

It fuses at 217° C, suddenly effervesces at higher temperatures, and suffers 
decomposition (see p. 214). 

Action of Hydrate of Barium on the Acid Sulphate. — The action which 
occurs when solutions of these two bodies are mixed varies in a remarkable 
manner with the conditions of the experiment. 



ACTION OF PHOSPHIDE OP SODIUM ON HALOID ETHERS. 197 

We naturally supposed that the hydrate of tetrabenzyl-phosphonium would 
be produced. Thus — 

(C 7 H 7 ) 4 PHS0 4 + Ba(OH) 2 = (C 7 H 7 ) 4 P(OH) + BaS0 4 + H 2 . 

In our first attempts to prepare the base by this method 3 grms. of the 
acid sulphate were dissolved in about half a litre of water, and a solution of 
caustic baryta (in boiling water) was added until no more sulphate of barium 
was precipitated. The solution was then filtered and a stream of carbonic acid 
passed through it until all excess of baryta was converted into carbonate. It 
was then boiled and filtered, and the filtered solution concentrated until it 
amounted to about 50 cc. in volume. Tufts of needles separated out, though 
in small quantity ; but on allowing the solution to stand for a couple of days, 
large rhombohedral plates appeared, which were colourless and of striking re- 
fractive power — resembling, in fact, iodide of phosphonium. They were washed 
with a little water and their mother liquor evaporated down, when a second 
crop of tabular crystals separated. The two crops were mixed and recrystal- 
lised from water. The resulting crystals were analysed. 

Heated in a drying oven to 110° C. they became opaque, and lost in weight, 
thus showing that they contained water of crystallisation. As, however, dur- 
ing drying they developed a smell, recalling that of bitter almonds, we cannot 
place much reliance on the water determination, as the compound no doubt 
volatilised slightly. It was, however, necessary to dry it for organic analysis, 
and we therefore heated it quickly in a small beaker placed in an oil bath until 
it had partially fused. As this only occurred at a temperature above 200° C, 
there can be no doubt that we obtained a dry product. 

Its combustion thus dried gave numbers agreeing with those required for 
the base. 

0-2065 grin, gave - 620 C0 2 = 81*8 per cent, carbon. 
0-2065 „ „ 0-1256 H 2 = 6-8 „ hydrogen. 

Obtained. Calculated for (C 7 H 7 ) 4 P(OH). 
Carbon, . . . 81-8 . . . 81-5 
Hydrogen, ... 6-8 .. . 7-0 

Wishing to prepare another quantity of the base, we proceeded in a similar 
manner, using however more concentrated solutions ; but on filtering and 
evaporating (after getting rid of excess of baryta), no base could be obtained. 
This curious result induced us to examine the precipitated sulphate of barium, 
which we found to contain a substance readily soluble in alcohol, though quite 
insoluble in water. It separated from its alcoholic solution in white needles. 
These were recrystallised until of constant melting point, viz., 210-212° C. 

VOL. XXX. PART I. 2 H 



198 PROFESSOR LETTS AND N. COLLIE ON THE 

They were then dried at 120°, but did not lose weight, so they were at once 
submitted to an elementary analysis. 

0-4234 gave 1-230 CO, = 792 per cent, carbon. 
0-4234 „ 0-2589 H 2 6 = 6-8 „ hydrogen. 

These results, so different from those required by the hydrate, we could not 
explain, and therefore again repeated the experiment with the acid sulphate 
and caustic baryta as far as possible under the same conditions as before — with 
larger quantities, however — but with exactly the same result. No body soluble 
in water was produced, but an insoluble body was precipitated with the 
sulphate of barium, soluble, however, as we had before noticed, in alcohol, and 
having on recrystallisation the same melting point (210-212° C). Its analysis 
gave numbers agreeing with those we had previously obtained. 

0-5958 gave 1-7221 C0 2 = 78-8 per cent, carbon. 
0-5958 „ 0-3636 H 2 = 6-8 „ hydrogen. 

Obviously then two distinct substances are formed by the action of caustic 
baryta on the acid sulphate, under conditions which vary only slightly from 
each other. 

On consideration, but one explanation appeared probable, viz., that the base 
suffered decomposition ; and the only change which appeared at all likely for 
it to undergo was a loss of hydrogen and benzyl, and, consequently, the forma- 
tion of oxide of tribenzyl-phosphine. This change would be analogous to the 
decomposition which Hofmann observed on heating the hydrate of tetrethyl- 
phosphonium, which he found to decompose with formation of ethane and 
oxide of triethyl-phosphine, 

(C 2 H 5 ) 4 P(OH) = (C 2 H 5 ) 3 PO + C 2 H 6 . 

Similarly, hydrate of tetrabenzyl-phosphonium might split up into toluol 
and the phosphine oxide, 

(C 7 H 7 ) 4 P(OH) = (C 7 H 7 ) 3 PO + C 7 H 8 . 

On comparing the percentages of hydrogen and carbon required for the 
oxide of tribenzyl-phosphine, with the results of our analyses of the body soluble 
in alcohol, we found that they corresponded. 

Obtained. Calculated for 

-i. " 5r (c 7 h 7 ) 3 po. 

Carbon, . . 79-2 . . 78-8 . . 7875 
Hydrogen, . . 6-8 6-8 . . 6-56 

It only remained to ascertain whether toluol was produced in the reaction. 
7 grms. of the acid sulphate were placed in a distilling flask, and about the same 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 199 

proportion of water as we employed in our two previous experiments. The 
two were then boiled until the sulphate had dissolved, and a boiling and con- 
centrated solution of caustic baryta was added in excess. The distilling flask 
was then connected with a condenser, and its contents heated. An oily liquid 
passed over with the water, which had the odour of toluol, and a boiling point 
of 112° — exactly that of toluol. Moreover, its weight amounted to TO grm., 
whereas the quantity we ought to have obtained was 1*2 grm. 

There can be no doubt then that the base very easily decomposes into 
toluol and tribenzyl-phosphine oxide. 

That this decomposition is effected by the temperature we must doubt, as 
we concentrated the solution of the base by boiling, and it suffered no change. 
But, on the other hand, there can be no doubt that if weak solutions of the 
acid sulphate and baryta are employed, the base is almost — if not altogether 
the sole product — whilst with strong solutions only the oxide is formed. 

Action of Carbonate of Barium on the Acid Sulphate. — Carbonate of barium, 
when added to a solution of the acid sulphate, at once effervesces, and sulphate 
of barium is formed, whilst the base remains in solution, from which it is easily 
obtained by evaporation and subsequent cooling. 

The reaction occurs according to the equation, 

(C 7 H 7 ) 4 PHS0 4 + BaC0 3 = C0 2 + BaS0 4 + (C 7 H 7 ) 4 P(OH) . 

We have not observed the production of any oxide of tribenzyl-phosphine 
in this reaction, even when a very concentrated solution of the acid sulphate is 
employed, and we therefore recommend it as the simplest and most certain 
method for obtaining the base, for the acid sulphate is very easily formed from 
the chloride. 

This reaction also shows that no carbonate of tetrabenzyl-phosphonium can 
exist ; nor ought we to be surprised at this, considering that the alkaline power 
of aromatic phosphines, like that of corresponding amines, is very markedly 
less than it is in the case of ethyl or methyl phosphines or amines. 

Hydrate of Tetrabenzyl-Phosphonium. — This body is one of the most beauti- 
ful of all the compounds of tetrabenzyl-phosphonium. It is very soluble even 
in cold water, and crystallises in very beautiful rhomboheclral plates, which 
we have on one occasion obtained nearly ^ an inch long and ^ inch broad. 
The crystals have a very striking refractive power, so that they resemble in 
appearance the large crystals of phosphonium iodide, which may be formed by 
slow sublimation. 



200 PROFESSOR LETTS AND N. COLLIE ON THE 

The hydrate is readily soluble in alcohol ; and the crystals obtained from the 
solution contain alcohol. 

0-8746 heated to 110° lost 01076 grm. alcohol = 12-3 per cent. 

Calculated for 
Obtained. (C 7 H 7 ) 4 P(OH),C 2 H 5 OH . 
Alcohol, . 12-3 12-8 

Solutions of the base have an alkaline reaction, and they neutralise acids — 
salts of the phosphonium resulting. 

When the anhydrous base is heated, it begins to fuse at 190°, but a portion 
remains unfused to 211°. It is probable, therefore, that the base cannot be 
melted without decomposition. Heated to higher temperatures it decomposes, 
and gives solid and liquid products, the nature of which we shall discuss later. 

Bromide of Tetrdbenzyl-Phosphonium. — This compound was obtained by 
mixing aqueous solutions of bromide of barium and the acid sulphate. The 
precipitated sulphate of barium contains the greater quantity of the bromide, 
which is sparingly soluble. It was extracted with alcohol, and the filtered 
solution mixed with water. On cooling, long silky needles were deposited. 
Their composition was verified by a bromine determination. 

•303 grm. required 6 - 3 cc. decinormal AgN0 3 soln. = 0*0504 Br = 16'6 per cent. 

Obtained. Calculated for (C 7 H 7 ) 4 PBr . 
Bromine, . 166 16-8 

The bromide is less soluble in water than the chloride. It melts at 216° to 
217°C. 

Iodide of Tetrabenzyl-Phosphonium was prepared by a similar process. It 
is almost insoluble in water, but is soluble in a mixture of water and alcohol, 
and crystallises from the solution in needles resembling the chloride. 

For its analysis, a quantity was fused, weighed, dissolved, and titrated 
with decinormal silver solution. 

•2103 grm. required 4 cc. AgN0 3 soln. = 0-0508 1 = 24-2 per cent. 

Obtained. Theory for (C 7 H 7 ) 4 PI . 
Iodine, . 24*2 24-3 

Nitrate of Tetrabenzyl-Phosphonium. — A solution of the base was saturated 
with a dilute solution of nitric acid ; long silky needles separated. They were 
not analysed. 

Chromate of Tetrabenzyl-Phosphonium. — This salt was obtained by acting on 
a solution of the chloride with chromate of silver. On extracting the precipitate 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHEES. 201 

with boiling alcohol, and adding water to the filtered solution, small lemon- 
yellow plates separated as the solution cooled. They were not analysed. 

Oxalate of Tetrabenzyl- Phosphonium. — On adding excess of oxalic acid to a 
dilute and hot solution of the base, and then allowing the mixture to cool, long 
needles separated. The salt was analysed by drying at 100° C, and determining 
oxalic acid in the dried compound by precipitation with chloride of calcium, 
the oxalate of calcium being eventually ignited and converted into quicklime. 

0741 gave 0-088 CaO = 0-1383 C 2 4 =17-9 per cent. 
0-770 lost 0-029 H 2 0= 3-7 



Obtained. 



Calculated for 
(C 7 H 7 ) 4 PHC 2 4 . 
C 2 4 . . 17-9 . . . 18-1 

{(C 7 H 7 ) 4 PHC 2 OJ,H 2 . 
H 2 . 3-7 3-6 

Acetate of Tetrabenzyl- Phosphonium. — This salt is the most soluble of any 
of the compounds of tetrabenzyl-phosphonium which we have examined. It is 
obtained by the action of the chloride on acetate of silver, and crystallises from 
alcohol with difficulty. It was not analysed. 

Chlorate of Tetrabenzyl-Phosphonium. — This salt was prepared by the action 
of chlorate of barium on the acid sulphate. It crystallises in long needles from 
a moderately concentrated solution. Heated above its melting point, it puffs. 

Examination of the Residues from the Preparation of Chloride of Tetrabenzyl- 
Phosphonium. 

After extracting the chloride of the phosphonium from the product of 
action of chloride of benzyl on phosphide of sodium (by boiling the latter 
repeatedly with water) a solid orange-coloured mass remains, which contains 
large quantities of amorphous phosphorus, and also organic phosphorus 
compounds. 

We thought it possible, if not indeed probable, that these would contain 
tribenzyl-phosphine, for it is only natural to suppose that that body is first 
formed in the reaction, and is only converted into the phosphonium compound 
by the continued action of the chloride of benzyl. If this were the case some 
of the phosphine might, we considered, remain unacted upon, and would be 
found in the residues. 

Moreover, we thought that phosphides of sodium of a different composition 
from the tri-sodium phosphide might be formed along with that body by the 



202 PROFESSOR LETTS AND N. COLLIE ON THE 

action of phosphorus on sodium ; for instance, a phosphide analogous in com- 
position to the liquid phosphide of hydrogen, viz., Na 4 P 2 , or even to the 
solid phosphide, viz., Na 2 P 4 . If these bodies were indeed formed, they might 
give by their action on chloride of benzyl analogous benzyl compounds, and of 
these the phosphorised cacodyl of benzyl, i.e. (C 7 H 7 ) 4 P 2 , would possess a 
great interest. 

We therefore determined to submit the residues to a very careful examina- 
tion, the result of which we have now to communicate. 

On boiling them with chloroform, the whole of the organic ingredients 
appeared to be dissolved, and the residue consisted apparently of amorphous 
phosphorus only. The chloroform extract was distilled to dryness, and there 
remained a solid gummy, slightly crystalline mass. This was boiled with 
alcohol, when the greater part dissolved, leaving about one-third undissolved 
in the form of a brown amorphous solid. The brown alcoholic extract was 
filtered and mixed with about one and a half times its volume of boiling 
water ; the mixture was then boiled and filtered from a brown resinous 
substance which had been precipitated. The filtered solution deposited on 
cooling a considerable quantity of a crystalline substance, though in a very 
impure condition. 

We thus split up the residues into three portions — 

1. Soluble in chloroform only. 

2. Soluble in chloroform and alcohol. 

3. Soluble in chloroform, alcohol, and a mixture of alcohol and water. 

We may here observe that as the portion of the residues No. 1 was not 
very inviting, in later experiments the residues were at once extracted with 
alcohol, so as to obtain only portions 2 and 3. 

Examination of Residue soluble in Chloroform, Alcohol, and Alcohol and 
Water. — This portion of the residue was distinctly crystalline, but highly 
charged with a brown resinous substance, which threatened at first to render 
its purification difficult. 

At first we attempted to recrystallise it from alcohol, but although well- 
defined crystals were easily obtained, the colouring matter adhered persistently; 
animal charcoal was then boiled with the alcoholic solution of some of the 
crystals which were only slightly coloured, and eventually by this process a 
colourless, crystalline, and highly refractive substance was obtained. 

But we soon found that the crude and strongly coloured residue could be 
purified in a far easier and simpler manner. On treating it with ether or 
bisulphide of carbon, the colouring matter was at once dissolved, but the 
crystalline substance was quite insoluble, so that by simple washing with either 
of these liquids, an almost colourless residue was obtained, which only required 
to be recrystallised once or twice from hot alcohol. The crystals thus obtained 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHEKS. 203 

consisted of brilliant needles of considerable length and thickness, and were 
apparently quadratic prisms. They fused at 212° C, and appeared to volatilise 
at higher temperatures almost unchanged. They were analysed by combustion 
with chromate of lead and oxide of copper. 

0-468 gave 1-360 C0 2 = 79"2 per cent, carbon. 
0-468 „ 0-2895 H,0 = 6-8 „ hydrogen. 

Carbon, .... 79-2 
Hydrogen, . . . 6 - 8 

That they contained phosphorus was at once shown by the livid green 
colour which they imparted to the flame of a Bunsen's burner when heated on 
a platinum wire. 

The numbers required for tribenzyl-phosphine do not agree with those 
obtained with this substance ; but, on the other hand, those calculated for a 
phosphorised cacodyl are almost identical with them. 

Thus— 







Calculated for 


Calculated for 




Obtained. 


(C 7 H T ) 4 P 2 . 


(C 7 H 7 ) 3 P. 


Carbon, 


. 79-2 . 


78-9 . 


82-9 


Hydrogen, 


. 6-8 . 


6-6 . 


6-9 



As for the reasons we have given above, we almost expected to find the 
cacodyl in the residues, we assumed that we really had obtained such a body. 
But the experiments which we tried to confirm this supposition did not lead to 
the results which we anticipated. 

A phosphorised cacodyl, even of an aromatic radical, ought to possess a 
strong tendency to absorb oxygen, and therefore would, even if it did not take 
up oxygen from the air, at least act as a powerful reducing agent. But the 
body in question showed no reducing tendencies whatever. A solution of 
nitrate of silver was not altered when boiled with its solution. Chloride of 
platinum at once produced a crystalline precipitate, which was not changed 
by boiling. This want of reducing power we considered a stong argument 
against the supposition that the substance was a phosphorised cacodyl. More- 
over, we noticed that it showed an indifference towards those reagents which 
react energetically with cacodyl. It is true that we obtained a compound 
with bromine, and also one with sulphur ; but the result of their analyses 
could not be reconciled with the supposition that they were derivatives of a 
phosphorised cacodyl. 

On looking over our notes, we were struck with the fact that, so far as 
carbon and hydrogen are concerned, both oxide of tribenzyl-phosphine and a 



204 PROFESSOR LETTS AND N. COLLIE ON THE 

phosphorised benzyl- cacodyl have an almost identical composition ; and we 
also noticed that dibenzyl-phosphine would require similar numbers. 

Thus — 

Calculated for 

(C 7 H T ) 4 P 2 . (C 7 H 7 ) 3 PO. (C 7 H 7 ) 2 HP. 
Carbon, . . 78-9 . . 797 . .78-9 
Hydrogen, . . 6*6 . . 6*6 . . 6'6 

We did not think it likely that dibenzyl-phosphine had been formed in the 
reaction, as we could not account for the hydrogen atom which it requires ; 
but bearing in mind the results of our experiments on the action of baryta on 
the acid sulphate of tetrabenzyl-phosphonium, it did not appear impossible 
that oxide of tribenzyl-phosphine had been formed ; for, by the action of phos- 
phide of sodium on water, caustic soda is produced : this might react on 
chloride of tetrabenzyl-phosphonium, and give rise to oxide of tribenzyl-phos- 
phine and toluol. 

At first sight, such a supposition may not appear probable, as haloid salts 
of methyl- and ethyl-phosphonium are not changed by caustic alkalies ; but we 
have shown that corresponding salts of benzyl-phosphonium possess very 
different £>roperties from these bodies. On treating the product of the action 
of phosphide of sodium on chloride of benzyl with water, abundance of phos- 
phuretted-hydrogen was evolved, showing that a considerable quantity of 
phosphide of sodium had remained unacted on. The solution was boiled ; and 
thus, if alkalies really act on chloride of tetrabenzyl-phosphonium in the manner 
we have indicated, we have the necessary conditions for the production of oxide 
of tribenzyl phosphine." 5 ' 

As a further argument for supposing that the oxide had really been 
obtained, and not the cacodyl, it will be noticed that, although the percentage 
of carbon calculated for the two bodies varies by only 0*8 per cent., the results 
of our analyses are more favourable to the supposition that the body is the 
oxide, and not the cacodyl. For we obtained 3 per cent, too much carbon 
for the cacodyl, and therefore 0*5 per cent, too little for the oxide ; and in 
carefully conducted organic analyses the carbon is often too low, but seldom 
too high. 

We had noticed that oxide of tribenzyl-phosphine (obtained as described 
at p. 198) combines with iodide of zinc to form a compound (analogous 
to Hofmann's zinc iodide compound of triethyl-phosphine oxide) of charac- 
teristic crystalline form. If, then, the substance were the phosphine oxide, the 

* We have since proved that alkalies act very readily on chloride of tetrabenzyl-phosphonium. On 
boiling a solution of the chloride in alcohol with potash or soda for a few minutes, decomposition occurs, 
chloride of the alkaline metal is precipitated and the solution contains oxide of tribenzyl-phosphine, 
which is easily identified by its insolubility in water and other characteristic properties. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 205 

production of this salt would be an almost crucial test. We therefore proceeded 
with the supposed cacodyl as we had done with the oxide of tribenzyl-phosphine, 
and operating under exactly the same conditions, obtained a double salt with 
zinc iodide, which could not be distinguished from that of the oxide, either in 
crystalline form or in composition. 

We have further verified the identity of the supposed cacodyl with oxide 
of tribenzyl-phosphine, by processes which we may consider along with the 
properties of that substance. 

Since writing the above, we have noticed that oxide of tribenzyl-phosphine 
has been obtained by F. Fleissner,"* by the action of benzal chloride on iodide 
of phosphonium. The results of Fleissner's investigations on the properties of 
the oxide, so far as they go, are in accordance with our own. 

Oxide of Tribenzyl-Phosphine. — Subjoined are the results of the analysis of 
the oxide prepared by three different methods : — 

I. Obtained as just described from the residues. 

II. and III. Obtained by the action of caustic baryta on chloride of tetra- 

benzyl-phosphonium. 
IV. Obtained during experiments on the action of sodium on chloride of 
tetrabenzyl-phosphonium (see p. 211). 







Obtained. 




Calculated for 
P(C 7 H 7 ) 3 

78-75 
6-56 

9-68 
5-01 


Carbon, . 
Hydrogen, 

Phosphorus, . 
Oxygen, 


I. 
. 79-2 
. 6-8 

. 8-5 8-8 


II. 

79-2 
6-8 




in. 

78-8 

6-8 


IV. 

78-3 
6-7 

8-4 




100-00 



The three specimens were quite different in appearance. 

I. Crystallised in thick needles of great refractive power, and quite trans- 
parent. 

II. and III. In opaque plates of indefinite form. 
IV. In very bulky, silky needles. 

We could not at first reconcile ourselves to the belief that they were one 
and the same body. 

The following carefully conducted experiments, however, appear to prove 
beyond doubt that they were so : — 

Melting Point. — This was determined in the ordinary manner, by heating the 
carefully dried and pulverised substance in a capillary tube tied to a thermo- 

* Ber. d. deutsch. chem. Ges., xiii., 1665. 
VOL. XXX. PART I. 2 I 



206 PROFESSOR LETTS AND N. COLLIE ON THE 

meter, both thermometer 4 " and capillary tube being placed in a beaker con- 
taining sulphuric acid. 

I. II. and III. IV. 

(a) 212° C. . ~212°^ . . . 212° 

(5) 212° . . . 212° . . . 210-212° 

The temperature is uncorrected. 

Double Salt ivith Zinc Iodide. — This compound was formed easily with 
any of the three specimens, by adding to its alcoholic solution an alcoholic 
solution of zinc iodide, and evaporating to small bulk. The double compound 
separates in thin transparent plates of characteristic form. 

Examined under the microscope no difference could be detected in the 
crystalline form of the double salt prepared with any of the three specimens of 
the oxide. 

The salt was analysed by volumetric determinationt of iodine in specimens 
of Nos. I. and of II. and III. — 

I. 0-606 grm. required 12*4 cc. decinormal AgNO 3 =0"15748= 26*0 per cent, iodine. 

II. and III. 0232 „ „ 4-7 „ „ „ =0-05969 = 25-9 

Calculated for {P(C 7 H 7 ) 3 0} 2 ZnI 2 26-4 

Chloroplatinate. — This salt is characteristic, and is formed with ease on 
mixing dilute alcoholic solutions of the oxide and chloride of platinum. It 
separates almost immediately as a light orange- coloured granular powder — 
which, under the microscope, is seen to consist of groups of needles, thick, 
four-sided, and with blunt ends. Very commonly two needles form a cross, 
at other times several radiate from a common centre. No difference could be 
detected in the crystalline form of the chloroplatinate prepared with any of the 
specimens. 

The salt was analysed by determination of carbon, hydrogen, and in one 
specimen of chlorine also : — 
I. 
Chlorine. 

04356 grm. required 15-3 cc. decinormal AgNO 3 =-054315 Cl = 12 - 4 per cent.f 
0-4605 „ „ 164 cc. „ „ =05822 „ =12-6 per cent.§ 

IV. 
Carbon and Hydrogen. 

(A) 0-3932 grm. gave 0*8497 C0 2 = 025491 C = 58-9 per cent. 
0-3932 „ „ 01874 H,0 = 002082 H = 5-3 



(B) 0-2738 „ „ 0-5975 C0 2 = 0-16295 C - 594 

0-2738 „ „ 0-1350 H 2 = 0-015 H = 5-4 „ 

* One of Casella's. 

t Volhardt's method. 

% Hofmann's method. 

§ By precipitating the platinum with sulphuretted hydrogen and titrating the filtered solution. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 207 

I. IV. 

Carbon, . - 58'9 _ 59 : 4 

Hydrogen, ...... 5*3 5-4 

Chlorine, 12-4 12-6 . . . 

Platinum, 



The formula which Hofmann "" gives for the chloroplatinate of the oxide of 
triethyl-phosphine is 3(Et 3 PO), Et 3 PCl 2 , PtCl 4 ; but this formula does not 
appear to be a very probable one. It seems to us to be more likely that 
the chloroplatinate is a compound of the phosphine oxide with hydrochloric 
acid and chloride of platinum, and we find that the numbers calculated for such 
a formula, viz., 4(Et 3 PO), 2HC1, PtCl 4 , agree as closely with those obtained by 
Hofmann in the analysis of the chloroplatinate, as do those calculated from his 

formula, thus — 

Calculated for 



Obtained. 3(Et 3 PO),(Et 3 PCl 2 ),PtCl 4 . 4(Et 3 PO),2HCl, PtCl 4 . 

Carbon, 3017 . . 30-9 .... 30-4 

Hydrogen, 675 . . . 6-4 .... 6-3 

Platinum, 21-06 . . . 21-2 .... 20-8 

Chlorine, 22-93 . . . 22-9 . . ... 22-5 

It will be seen that the only difference between these two formulae is that 
the one on the right hand contains an atom more oxygen and two atoms more 
hydrogen than the one on the left ; that is to say, a difference of 18 as regards 
molecular weight. As the latter amounts to 930 in Hofmann's formula, the differ- 
ence in the calculated percentage of each element is very slight, and this is still 
more the case with the chloroplatinate of the benzyl compound — the molecular 
weight of which with Hofmann's formula is 1674, and with our formula 1692. 

But, on the other hand, the proportion of carbon is so large that the percen- 
tage of that element is perceptibly different with the two formulae, and it will 
be seen that this difference is decidedly in favour of the formula which we 
propose. 

We may add that it appears to us to be highly improbable that O should be 
replaced by Cl 2 , by simply mixing at ordinary temperatures chloride of plati- 
num and the phosphine oxide.t 

The results of our analysis, compared with the numbers calculated for the 

two formulae are — 

Calculated for Calculated for 



Obtained. 
Carbon, 59'2 


3{(C 7 H 7 ) 3 PO}, {(C 7 H 7 ) 3 PC1 2 |, PtCl 4 
60-2 


4{(C 7 H 7 ) 3 PO}2HCl,PtCl 4 . 
59-5 


Hydrogen, 5-3 


5-0 


5-0 


Platinum, 


11-7 


11-7 


Chlorine, 12'5 


12-7 


12-5 



* Trans. Eoy. Soc. Lond., I860, p. 418. 

t The experiments of Crafts and Silva (he. cit.) show that this replacement does not occur. 



208 PROFESSOR LETTS AND N. COLLIE ON THE 

Brominated Compound. — This is a very characteristic substance, and its 
production, with all of the specimens of the supposed oxide, we considered to 
be a strong proof of their identity. 

It is formed by dissolving the phosphine oxide in glacial acetic acid (boiling), 
and adding bromine in excess. No visible reaction occurs, except that the 
bromine is at first decolorised. On cooling, the compound is precipitated as a 
granular crystalline powder of bright yellow colour. Sometimes needles are 
observed ; but these are found, when examined under the microscope, to con- 
sist of cubical or rhombohedral crystals united ; the crystalline powder consist- 
ing of the same forms isolated. 

For analysis, the compound was simply dried in vacuo for some time, and 
was not recrystallised. 

Carbon and Hydrogen. * 

0-4746 grm. gave 0-9915 C0 2 = 0-2704C = 56-9 per cent. 

0-4746 „ „ 0-2117 H 2 = 0-0235H = 4-9 „ 
Phosphorus, -f- 

0-6777 required 16-1 cc. uranium solution = 0-0368 P = 5-2 per cent. 
Bromine. \ 

0-1685 required 6 - cc. silver solution = 0'048 Br = 284 per cent. 

0-2128 „ 7-55 cc. „ „ = 0-060 Br = 28-3 

0-3498 „ 12-37 cc. „ „ = 0-099 Br = 28-3 

These numbers agree closely with the rather curious formula, 

4 {(C 7 H 7 ) 3 POBr 2 K(C 7 H 7 ) 3 PO, or 5{(C 7 H 7 ) 3 PO}, 4Br 2 , 
but with no other that appeared probable. 

Obtained. Calculated. 





I. 


II. 


in. 




Carbon, 


56-9 


— 


— 


56-3 


Hydrogen, . 


49 


— 


— 


4-7 


Phosphorus, 


5-2 


— 


— 


6-9 


Bromine, 


284 


28-3 


28-3 


28-5 



The bromine compound when treated with acetic acid loses bromine. It 
cannot, therefore, be readily recrystallised. Heated by itself it fuses, but at no 
definite temperature, to a deep yellow liquid. Hydrobromic acid is then given 
off, and later bromide of benzyl (?) distills. Heated with alcohol it dissolves, 
and the solution (at first yellow) gradually becomes colourless, and the odour of 
bromide of benzyl is apparent ; but a considerable quantity of bromine may be 
precipitated by nitrate of silver from the alcoholic solution. When boiled with 
water it decomposes, and bromine is evolved. 

* By combustion with oxide of copper and cbromate of lead. 

t Fused in a silver dish with caustic potash and nitrate of potash, and subsequently titrated with 
uranium solution. 

X Fused in a silver dish with caustic potash, and subsequently titrated by Volhardt's method. 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 209 

Sulphuretted Compound. — When the phosphine oxide is fused with sulphur 
a reaction occurs, which apparently varies with the temperature and with 
the quantity of sulphur employed. If much sulphur is taken and the mixture 
heated to a high temperature, sulphuretted hydrogen is evolved, the mass 
becomes dark coloured, and resinous products are formed. 

But if the proportion of sulphur is low (P(C 7 H 7 ) 3 : S. 2 ) and the temperature 
is kept at the melting point of the oxide or rather higher (240°), the sulphur 
dissolves, no gas is evolved, and the product dissolves completely in a large 
quantity of boiling alcohol. The solution on cooling deposits beautiful silky 
needles of a light buff colour, which do not readily change in appearance (nor 
alter in their melting point) by recrystallisation. That the new substance 
contains sulphur is shown by burning it on platinum foil, when a strong odour 
of sulphurous anhydride is at once observed. 

The substance fuses at 211°-212° (uncorrected). It is insoluble in water, 
and sparingly soluble in alcohol. The only determinations made were of the 
carbon and hydrogen which it contains. 

0-2103 gave 0-597 C0 9 =0-16254 C = 77"3 per cent. 
0-2103 „ 0-132 H 2 6 = 0-01466 H= 6-9 

These numbers do not agree with any simple addition product. The only 
probable formula which agrees with the numbers obtained is, 

4 {(C 7 H 7 ) 3 PO| , (C 7 H 7 ) 3 POS = 5 {(C 7 H 7 ) 3 PO} , S . 

Thus— 





Obtained. 


Calculated 


Carbon, 


77-3 


77-3 


Hydrogen, . 


6-9 


6-4 



Examination of Residue, soluble in Chloroform and Alcohol only. 

This portion of the residue was contained in the dark brown mother liquors 
of the crystalline substance, which the preceding experiments have shown was 
oxide of tribenzyl-phosphine. 

On evaporating off the alcohol a dark brown syrupy mass remained, which 
solidified on cooling to a resin. This contained some crystalline matter, which 
we could not succeed in separating. We have subjected the resin to many 
experiments with the view of obtaining definite products, only, however, with 
partial success. 

In one of our earlier experiments we subjected it to the action of heat. 55 
grms. were heated in a distilling flask. The thermometer rose rapidly to 
380°, and a small quantity of a solid substance distilled. The temperature 
then fell suddenly, and a liquid distillate was obtained. After some time the 
temperature again rose above the boiling point of mercury, and the residue 
began to char. The products of this distillation were collected together and 



210 PROFESSOR LETTS AND N. COLLIE ON THE 

redistilled. They began to boil a little above 100° C. The distillate was 
divided into two fractions, viz., from 100°-200° C., and from 200°-320° C. 

The first of these was liquid, and on redistillation passed almost entirely 
between 1 10°-114° C. (chiefly at 112° C), and had all the properties of toluol. The 
second was solid, and contained a large quantity of free phosphorus. As 
its fractional distillation did not give satisfactory results it was dissolved in 
boiling alcohol. Free phosphorus in some quantity remained undissolved, 
and on filtering and cooling the solution, colourless crystals separated. They 
were collected and recrystallised until their melting point was constant, viz., 
118° C. 

This is the melting point given by Laurent for stilbene, and the crystalline 
habit which is so characteristic was exactly the same as that of the substance 
under examination. On combustion we obtained numbers agreeing fairly well 
with those calculated for that hydrocarbon. 

0*3135 gave 1-0815 carbonic anhydride =0'29495 carbon = 94-1 per cent. 
0-3135 „ 0-1965 water =0-02183 hydrogen=6-9 





Obtained. 


Calculated 


Carbon, 


94-1 


93-3 


Hydrogen, . 


6-9 


6-6 



The mother liquors from which it had been separated were concentrated, and 
yielded a batch of colourless crystals, which were not examined. The mother 
liquors from them were considerably concentrated, and yielded another crop of 
colourless crystals, which, after repeated recrystallisation, ceased to alter in 
melting point. This was 51° C, which is that of dibenzyl. We have not 
analysed the substance, as we considered its identity with dibenzyl proved by 
its melting point and characteristic odour. 

We had thus proved that the resin split up on heating into free phosphorus, 
stilbene, dibenzyl, and toluol — a result which might occur supposing it to 
have consisted of tribenzyl-phosphine. The equation, 

Stilbene. Dibenzyl. Toluol. 

2(C 7 H 7 ) 3 P = 2P + C^ 2 + C^ 4 + 2C 7 H 8 . 

shows this. 

This supposition is strengthened by the fact that sulphide of benzyl yields 
stilbene when heated ; and one of us has shown that organic compounds of 
phosphorus and sulphur often behave in a similar manner. 

Moreover, subsequent experiments showed that chloride of tetrabenzyl- 
phosphonium is decomposed by heat into hydrochloric acid, and the same pro- 
ducts as we obtained on heating the resin. We also heated the resin with 
chloride of benzyl in a sealed tube for some time at 180°-190°C. Nothing par- 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHEPS. 211 

ticular appeared to have occurred, the contents of the tube consisting of a brown 
viscous mass. But on boiling this with water, and cooling the solution, chloride 
of tetrabenzyl-phosphonium crystallised out, and was proved to be pure by a 
determination of chlorine. 

0656 required 13"5 ec. decinormal AgN0 3 = 7 , 3 per cent. 01. 
(C 7 H 7 ) 4 PC1 ; 2H 2 requires . . .7*6 

This experiment would have definitely proved the resin to consist of tribenzyl- 
phosphine, had the phosphonium chloride been produced in large quantity ; but 
such was not the case, for about 20 grms. of resin only gave about 2 grms. of 
the chloride ; still it shows that the resin contained the phosphine. 

As Hofmann (Joe. cit.) has found that dibenzyl-phosphine does not combine 
with acids, we could scarcely expect to obtain salts of the tertiary-phosphine. 
We, however, heated the resin with fuming hydrochloric acid, but, as we 
expected, obtained no salt. We have also tried to obtain the well characterised 
oxide of tribenzyl-phosphine, by treating the resin with various oxidising 
agents, but without success. Nor could we obtain any definite compounds on 
treating the resin with bromine, chloride of platinum, or iodide of zinc. We 
therefore abandoned its further investigation. 

Attempts to prepare Tribenzyl-Phosphine. 

So far as we are aware, no method has been discovered for converting 
the oxide of a tertiary-phosphine or a salt of a compound phosphonium into a 
tertiary-phosphine itself. 

Considering the remarkable stability of the former class of bodies, and the 
tenacity with which the oxygen adheres to the phosphorus, we scarcely expected 
to effect the reduction of the oxide of tribenzyl-phosphine. We, however, sub- 
jected it to the action of potassium, of sodium, and of zinc dust, but, as we 
expected, it either remained unchanged, or suffered complete decomposition. 

We hoped, however, to meet with better success in attempting to obtain 
tribenzyl-phosphine from chloride of tetrabenzyl-phosphonium. Two methods 
suggested themselves to us, the first being to act on the chloride with sodium, 
which we anticipated would give chloride of sodium, dibenzyl, and the 
phosphine, 

2[(C 7 H 7 ) 4 PC1] + Na 2 = 2NaCl + C 14 H U + 2(C 7 H 7 ) 3 P . 

A preliminary experiment showed that when chloride of tetrabenzyl- 
phosphonium is boiled for some hours with xylol and sodium, chloride of 
sodium is produced. 

We therefore made a carefully conducted experiment as follows : — 

24 grms. of the pure chloride were carefully dried and introduced into a 



212 PROFESSOR LETTS AND N. COLLIE ON THE 

flaek connected with a reversed condenser. 100 grins, of redistilled xylol 
(boiling point 135°-137° C.) were then added together with l - 3 grm. of sodium. 
A current of hydrogen was then passed through the apparatus, and the mixture 
kept gently boiling. When most of the sodium had been acted on (which 
required some days' digestion), the xylol was poured off and filtered. On 
cooling, it deposited an abundance of crystalline matter. This was collected 
on a cloth filter, well squeezed to free it from adhering xylol, and dissolved in 
boiling alcohol. On cooling, crystals separated having the appearance of oxide 
of tribenzyl-phosphine, and which were proved to consist of that body. The 
xylol from which this oxide had separated was distilled to dryness, and the 
residue taken up with boiling alcohol. The solution on cooling deposited beau- 
tiful silky needles, which were recrystallised twice from alcohol. In spite of 
their very different appearance from other specimens of oxide of tribenzyl- 
phosphine, a most careful examination showed that they consisted of that body 
(see p. 205). We are completely unable to account for the difference in appear- 
ance of the two quantities of the oxide obtained in this experiment. No one 
would imagine that they were the same body. We could not obtain any other 
definite products from this experiment. 

Now the production of the oxide may be accounted for in two ways — (1) 
the chloride of tetrabenzyl phosphonium was not perfectly dry, and caustic 
soda was formed, which then acted upon it (as we have already shown), to give 
toluol, common salt, and the oxide ; (2) tribenzyl-phosphine was formed, and 
absorbed oxygen from the air during the subsequent processes to which the pro- 
duct of the reaction was submitted. We have repeated the experiment several 
times, using every precaution to prevent water or oxygen from coming in con- 
tact with the mixture of sodium, xylol, and the phosphonium chloride, but 
always with the same result — viz., production of large quantities of the oxide. 
At present we do not know which of the two explanations we have given of its 
production is the correct one. 

We may mention that finely divided silver acts on the chloride of tetra- 
benzyl-phosphonium when the two are heated together ; the action, however, 
only occurs to a slight extent, and we were not successful in obtaining any 
definite product. 

The other method that occurred to us for obtaining tribenzyl-phosphine 
from the chloride of tetrabenzyl-phosphonium was to treat the latter with 
phosphide of sodium, which we hoped would react so as to give tribenzyl 
phosphine and chloride of sodium, 

3(C 7 H 7 ) 4 PC1 + Na 3 P - 3NaCl + 4(C 7 H 7 ) 3 P . 

The following experiment was made : — 3 grms. of phosphide of sodium and a 
little xylol were heated in a sealed tube, with 13 grms. of chloride of tetrabenzyl- 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 213 

phosphonium. After three days' heating at a temperature of 180°-190° most of 
the phosphide was acted on. The tube was then opened and repeatedly extracted 
with dry chloroform. The extract was distilled to dryness, and the residue 
treated with ether, which dissolved most of it, but left a small quantity of the 
oxide of tribenzyl-phosphine. The ethereal extract was evaporated to dry- 
ness, and left a light-coloured soft resin, which partly crystallised. A phos- 
phorus determination showed that this contained the quantity of that element 
calculated for tribenzyl phosphine, 

0"651 required* 29 - l cc. uranium solution = 9 - 9 per cent. P 
<C 7 H r ) s P requires 10-2 

The quantity, however, of the resin at our disposal was so small that we could 
not make a thorough investigation of it. But we are inclined to the belief that 
both it and the resin obtained as a bye product in the preparation of the chloride 
of tetrabenzyl-phosphonium consisted mainly of tribenzyl-phosphine (see p. 209). 

Action of Heat on the Salts of Tetrabenzyl-Phosphonium. 
During the experiments we have described, we obtained on heating several 
of the salts of tetrabenzyl-phosphonium, results which invited a closer investiga- 
tion. Partly on this account, and partly from the interesting results which Drs 
Crum Brown and Blaikie t have obtained by the action of heat on the salts of 
trimethyl-sulphine, we determined to study the behaviour of one or two of the 
compounds of tetrabenzyl-phosphonium when heated. 

Action of Heat on Chloride of Tetrabenzyl-Phosphonium. — We hoped that 
the salt would dissociate when submitted to the action of heat into chloride of 
benzyl and tribenzyl-phosphine. 

A considerable quantity of the chloride previously dried and fused was 
placed in a small distilling flask and heated in an air bath. Nothing particular 
occurred until the temperature had risen to about 300° C, when the fused salt 
began to grow brown, and a colourless liquid distilled. When a considerable 
quantity of this had passed over, hydrochloric acid was evolved, and later the 
distillate was yellow, and contained an abundance of free phosphorus. The 
heating was continued until nothing further distilled. There remained a con- 
siderable residue, consisting chiefly of charcoal. 

The whole of the distillate was fractionated. Hydrochloric acid was evolved 
in abundance; the thermometer then rose to 109°, and by far the larger quantity 
of the product passed over between that temperature and 115°. This fraction 
on redistillation boiled constantly at 110°-113°, and had the odour of toluol. 
It was not further examined, and was considered to be that substance. 

* After fusion with, a mixture of nitrate of potash and caustic potash. 
\ Proceedings Roy. Soc, Edin. 

VOL. XXX. PART I. 2 K 



214 PROFESSOR LETTS AND N. COLLIE ON THE 

The higher boiling residue passed between 280°- 300°, and solidified in the 
condenser. It was dissolved in alcohol, and recrystallised several times. The 
recrystallised substance had the characteristic form and melting point (118° C.) 
of stilbene. 

In the mother liquors there remained a solid of lower melting point, and 
having the odour of dibenzyl ; but its quantity was too small to enable us to 
identify it absolutely. We think that there can be but little doubt that 
it consisted of that body. 

No chloride of benzyl could be found, although the liquid product certainly 
smelt of it. Its quantity was therefore insignificant. 

This experiment shows that the phosphonium chloride is not dissociated by 
heat, but splits up in a somewhat complicated manner. Very possibly the first 
action of heat is to give stilbene, hydrochloric acid, and tribenzyl-phosphine. 



,(^ 7 1 3 CH;H ^P-Cl) = 2(C 7 H 7 ) 3 =P + C 6 H 5 -CH = CH-C 6 H 5 4 
The phosphine splitting up later into toluol, stilbene, and dibenzyl 



/ C .6. H 5- CH i H \ \ C 6 H 5 -CH=CH-C 6 H 5 + C 6 H 5 -CH 3 

2(aH 5 -CHH_\p] = 



Vc^-CHHi/^ / +C 6 H 5 -CH 2 -CH 2 -C 6 H 5 + 2P. 

It is however quite possible, considering the large quantity of toluol which 
is formed in proportion to the stilbene, and also considering the considerable 
amount of charred matter which remains, that the tribenzyl-phosphine splits up 



into toluol, and the residue C 6 H 5 — C, only, the latter becoming carbonised 



C 6 H 5 — CH 2 



C 6 H 5 - C = 



+ 2C r H s -CH 



6 ix 5 



Action of Heat on the Acid Sulphate. — 8 grms. of the acid sulphate were 
carefully dried, and heated in a small retort connected with a wide con- 
densing tube. The salt fused, then effervesced violently, and a colourless liquid 
distilled which solidified in the condenser. Sulphurous anhydride was given 
off at the end of the operation, and a slight residue of syrupy consistency 
and of a dark brown colour remained in the retort. The crystalline product 
was washed with alcohol until quite colourless, and then recrystallised several 
times from the same liquid, in which it was not very soluble. It crystallised 
in very thin needles of considerable length. These melted at 205°-206° C. 
It did not precipitate chloride of barium, but contained sulphur, as it gave 
the sulphuric acid reaction after it had been oxidised with a mixture of 



ACTION OF PHOSPHIDE OF SODIUM ON HALOID ETHERS. 215 

nitric acid and chlorate of potash ; and molybdate of ammonia showed that 
phosphoric acid was also present in the substance thus oxidised. No chloro- 
platinate could be obtained, but on mixing alcoholic solutions of the substance 
and of chloride of platinum a black precipitate was produced, consisting either 
of reduced platinum or of its sulphide. 

The substance was burnt with chromate of lead and oxide of copper, and 
gave numbers agreeing with those required for the sulphide of tribenzyl- 
phosphine. 

0-3567 gave 1-0097 C0 4 = 0-27587 C = 74-4 per cent. 



0-3567 „ 0-2114 H 2 = 0-02348 H= 6-6 





Obtained. 


Carbon, 


74-4 


Hydrogen, . 


6-6 



Calculated for 
(C 7 H 7 ) 3 PS. 
75-0 
6-2 



The compound was not further examined. 



Action of Heat on the Hydrate. — From the experiments described at p. 196 
on the action of caustic baryta on the acid sulphate, we were led to think that 
the latter would easily split up into toluol and oxide of tribenzyl-phosphine, 
and we therefore determined to ascertain if this supposition were correct. 

A quantity of the hydrate crystallised from alcohol was placed in a distilling 
flask and heated in an oil bath. The alcohol of crystallisation first passed off, 
and at 250° C. the compound melted, and immediately a colourless liquid began 
to distil, which ceased to pass over at 260° C. The liquid was redistilled and 
boiled constantly at 111°-112° C. It consisted therefore of toluol. 

The residue in the distilling flask crystallised on cooling, was insoluble in 
water (whereas the hydrate readily dissolves), but was soluble in alcohol, and 
crystallised in the characteristic form of the oxide. Its melting point was 
found to be 212° C, and it gave the characteristic brominated compound and 
chloroplatinate of the oxide. 

The decomposition which the hydrate suffers when heated may therefore be 
expressed by the equation — 

(C 7 H 7 ) 4 P(OH) = (C 7 H 7 ) 3 PO + C 7 H 8 . 

Action of Heat on Oxide of Tribenzyl-Phosphine. — The oxide partly volatilises 
unchanged when it is heated, and partly decomposes into toluol, free phosphorus, 
charred matters, and other substances obtained in too small quantity to be 
investigated. 



( 217 ) 



IX. —On the Geology of the Fcerbe Islands. By James Geikie, LL.D., 
F.R.S. L. & E. (Plates XIII., XIV., XV., XVI.) 

(Read March 15, 1880.) 



CONTENTS. 



I. Introduction 

II. Physical Features 
Islands. 



PAGE 

218 



OF THE 



1. Extent, Form, and Trend of 

the Islands and Fiords . 220 



Configuration and Height of 

the Land 
Valleys 
Lakes and Streams 



III. Geological Structure op 
Islands. 



the 



1. General Dip of the Strata 

2. Contemporaneous or Bedded 

Basalt-rocks of Suderoe 

3. Bedded Tuffs of Suderoe 

4. Coal and Coal-bearing Beds of 

Suderoe 

5. Coal, &c. of Myggenaes and 

Tindholm 

6. Subsequent or Intrusive Basalts 

of Suderoe . 

7. Contemporaneous or Bedded 

Basalt-rocks of Northern 
Islands 

8. Bedded Tuffs of Northern 

Islands 

9. Subsequent or Intrusive Basalts 

of Northern Islands . 

IV. Thickness op the Strata : Con- 
ditions UNDER WHICH THEY 
were Amassed. 

1. Thickness of the Strata 

2. Igneous Rocks of Subaerial 

Origin 

VOL. XXX. PART I. 



221 
222 

222 



223 

223 

226 

227 
229 
230 

231 
235 
235 



237 
237 



3. Miocene Age of the Strata : 

Physical Conditions, etc. 

4. Position of old Volcanic Centre 



V Glacial Phenomena 
Islands. 



of THE 



1. Early Notices of Glacial Phe- 

nomena 

2. Glaciation 

3. Till or Boulder-clay 

4. Erratics and Morainic Debris 

5. Lake-Basins . 

VI. Origin op the Valleys and Fiords 
Subaerial and Glacial Ero- 
sion. 

1. Forms of Valleys 

2. Fiords 

3. Trend of Valleys and Fiords : 

Main Water-parting 

4. Origin of Main Water-parting 

5. Atmospheric Erosion 

6. Former Greater Bainfall 

7. Glacial Erosion of Valleys 

8. Weathering of Glaciated Sur- 

face 

9. Limited Accumulation of Till 

on Land 

10. Direction of Ice-flow and 

Extent of Ice-sheet . 

1 1 . Origin of Erratics and Morainic 

Debris 

VII. Marine Erosion 
VIII. Peat and Buried Trees 
IX. Explanation of Plates 

2l 



240 
242 



243 
244 
249 
250 
251 



253 
254 

255 
256 
257 
259 
260 

260 

260 

261 

262 
263 
266 

267 



218 DR JAMES GEIKIE ON 



I. Introduction. 

In this paper I give an account of observations made during a visit in 1879 
to the Freroe Islands in company with my friend Mr Amund Helland of 
Christiania. The principal object of our journey was to examine the glacial 
phenomena of the islands, but we studied so far as we could the various rock- 
masses of which the group is composed, and constructed a geological sketch- 
map to show the line of outcrop of coal, the disposition of the strata, the 
direction of dykes, and the trend of the glaciation. I have only to add, that 
all the observations recorded in the following pages were made in concert with 
my friend, and I am glad to say that we were quite at one in our general 
conclusions.* 

The earliest references to the geology of the Fseroe Islands are met with in 
a general description of the group by Lucas Jacobson DEBEst (1673), but, as 
might have been expected, they are of no scientific value. He makes brief 
reference to the occurrence of coal in Suderoe, stating that it is found in only 
one place " to which one can with difficulty come ; " from which it is probable 
that he had in view some of the outcrops in the precipitous sea-cliffs. 

In 1800 appeared a general account of the islands by Jorgen Landt, a 
resident Danish clergyman, in which the physical features of the islands are 
well described.^ The author was no geologist, but he notes some of the more 
characteristic aspects of the rocks, and was clearly of opinion that some of 
these at least had been in a state of fusion. He also gives some account of 
the many large angular perched blocks which are so plentifully sprinkled over 
the islands. It was Landt's description of the igneous rocks which induced 
Sir George Mackenzie to visit the islands. Sir George was accompanied by 
Mr Thomas Allan, and each subsequently gave an account of his own obser- 
vations ; the papers appearing in an early volume of the Transactions of this 
Society.§ 

Sir George Mackenzie limits his remarks on the " trap " of the Faeroes to 
such characters as seemed to him to demonstrate the igneous origin of that 
class of rocks. He distinguishes between the "tuff" or "tuffa" and the 
" trap ; " shows how they are interbedded ; and gives the general inclination of 

* Mr Helland's paper has been published since the present memoir was read. See "Om Fsero- 
ernes Geologi," in the Danish Geografisk Tidskrift, 1881. 

t Fserose et Fseroa Eeserata, &c., Kiobenhafn, 1673. 

X Forsog til en Beskrivclse over Faeroerne, Kiobenhavn, 1800. An English translation of Landt's 
work appeared in 1810. 

§ "An Account of some Geological Facts observed in the Faroe Islands" (Mackenzie), Trans. 
Roy. Soc. Edin., vol. vii. p. 213 ; and "An Account of the Mineralogy of the Faroe Islands " (Allan), 
"//. tit. p. 229. 



THE GEOLOGY OF THE F^EROE ISLANDS. 219 

the strata as towards south-east at an angle of about 4° or 5°. He was of 
opinion that the bedded traps had been ejected from submarine volcanoes. 

Mr Allan's paper is chiefly mineralogical, but he also gives some geological 
details. Both he and Mackenzie noticed the dykes that here and there inter- 
sect the strata, but only Mr Allan describes the irregular masses of " green- 
stone " which are unconformable to the regular bedded trappean rocks among 
which they occur. He also insists upon the igneous formation of all the traps, 
but does not commit himself to Mackenzie's submarine-volcano theory. The 
circumstances under which the traps were formed seem to him as inexplicable 
as ever, but he evidently leans to the view of their subaerial origin. He 
describes the smoothed appearance of the sides of the mountains, and particu- 
larly refers to a place at Eide in Osteroe where " the rock is scooped and 
scratched in a very wonderful degree, not only on the horizontal surface, but 
also on a vertical one of 30 to 40 feet high, which had been opposed to the 
current, and presented the same scooped and polished appearance with the 
rest of the rock, both above and below." These phenomena he recognises to 
be the same as the smoothed and dressed rocks near Edinburgh. 

Mackenzie's and Allan's papers were supplemented by Mr W. C Trevel- 
yan, who, in a letter to Dr Brewster,* gives descriptions of the geology of 
Myggenaes and Suderoe — two of the islands which Mackenzie and Allan 
were unable to visit. His short description of the coal-beds of Suderoe is 
correct so far as it goes, but, curiously enough, he says the beds dip south- 
east, while the section given by him shows them dipping to the north. The 
letter is accompanied by some excellent sketch-sections, exhibiting the 
appearances presented by certain irregular masses of basalt. 

A few years later Dr Forchhammer, who does not appear to have known of 
Mackenzie's and Allan's papers, visited the islands at the instance of the 
Danish Government, and afterwards published a very able description of their 
geognosy ,t accompanied by an admirable geological map. His observations 
and views, however, I shall refer to more particularly in the sequel. He makes 
no reference to the phenomena of smoothed rocks which so impressed Allan. 

The next geological notice of the Fseroe Islands occurs in a series of articles 
by Robert Chambers, entitled " Tracings in Iceland and the Fgeroe Islands." X 
He spent only some two or three days among the group, but recognised marks 
of glaciation at various places, as I shall afterwards point out. 

Since the date of his visit, the islands have frequently been referred to in 
books of travel, but none of these has added anything to what was already 

* "On the Mineralogy of the Faroe Islands" Trans. Roy. Soc. Edin., vol. ix. p. 461. 
f " Om Freroernes geognostiske Beskaffenhed," Det kongl. danske Vidensk. Selsk. Skrifter, 1824. 
See also Karsten's Archiv. fiir Mineralogie; vol. ii. p. 197. 



220 . DR JAMES GEIKIE ON 

known. In 1873, however, appeared an excellent paper by Professor John- 
strup, in which he gives a detailed account of the coal-beds of Suderoe. # 
This, I believe, is the most recent addition to our knowledge of the geology of 
the Fseroe Islands. It is referred to in my description of Suderoe. t 

II. Physical Features of the Islands. 

1. Extent, Form, and Trend of the Islands and Fiords. — The Fserde Islands \ 
are upwards of twenty in number, and nearly all are inhabited. They extend 
over an area of about seventy miles in length from north to south, and nearly 
fifty miles in breadth from west to east. The two largest islands are Stromoe 
and Osteroe, which closely adjoin and contain together upwards of 250 square 
miles, an area which is nearly equal to that of all the other members of the 
group. The extent of land in this little archipelago may therefore be roughly 
estimated at about 500 square miles. Nearly all the islands have an elongated 
form, and are drawn out in a N.N.W. and S.S.E. direction. This is the direction 
also of the more or less narrow sounds or open fiords that separate the islands 
in the northern part of the archipelago ; and the wider belts of water in the 
south, such as Suderoe Fiord, Skuoe Fiord, and Skaapen Fiord, have the same 
general trend. A glance at the accompanying map (Plate XVI.) will show that 
many of the closed fiords which penetrate the islands extend in a similar direc- 
tion throughout the whole or a large part of their course. There are no great 
depths in the immediate vicinity of the islands. None of the closed fiords is so 
deep as many of the Scottish and Norwegian sea-lochs, the deepest soundings 
indicated upon the charts never exceeding 65 fathoms. The soundings, how- 
ever, are few in number, and we were told by the fishermen of considerably 
greater depths in some places than are shown on the chart. Thus we were 
assured that Skaalefiord is 40 or 50 fathoms deep. Immediately outside of the 
islands the sea-bottom appears to slope away somewhat gradually in all 
directions until a depth of upwards of 100 fathoms is reached at a distance of 
15 or 20 miles, more or less, from the nearest coast-line. 

* "Om Kullagene paa Faerberne samt Analyser af de i Danmark og de nordiske Bilande forekom- 
mende Kul," K. D. Vidsk. Selsk. Oversigt, 1873, p. 147. 

t Since the above was written, I have met with another paper referring to Suderoe, by A. H. Stokes, 
ELM. Inspector of Mines, in "Trans. Chesterfield and Derbyshire Institute of Mining, Civil, and 
Mechanical Engineers," vol. ii. p. 320. The author seems to have examined only the mines and outcrops 
in the Trangjisvaag district, and he gives the average thickness of the coal seen by him, together with 
the heights above the seadevel of the various points at which the seam crops. He gives also analyses of 
the coal. He upholds the submarine origin of the volcanic rocks, and thinks the coal may be the 
remains of driftwood floated from America. 

\ For the spelling of place-names, I have followed the Danish Chart, although the orthography 
differs from that used in other Danish works. Some of the places I refer to are not given on the chart, 
and for the spelling of these I am indebted to my colleague Mr Helland. A number of the place- 
names in Suderoe, I have taken from the map accompanying Professor Johnstrup's paper. 



THE GEOLOGY OE THE FJEROE ISLANDS. 221 

2. Configuration and Height of the Land. — The islands are for the most part 
high and steep, many of them being mere narrow mountain-ridges that sink 
abruptly on one or both sides into the sea. The larger ones, such as Stromoe, 
Osteroe, and Suderoe, show more diversity of surface, but they possess very 
little level land. All the islands have a mountainous character — the hills, 
owing to the similarity of their geological structure, exhibiting little variety of 
feature. These high grounds form as a rule straggling, irregular, flat-topped 
masses, and sharper ridges which are notched or broken here and there into a 
series of more or less isolated peaks and truncated pyramids. Sometimes the 
mountains rise in gentle acclivities, but more generally they show steep and 
abrupt slopes, which in several instances are found to have inclinations of 25° 
to 27°, and even 30°. In many places they are still steeper, their upper 
portions especially becoming quite precipitous. They everywhere exhibit the 
well-known terraced character which is so common a feature of trappean masses. 
Precipices and long cliffs or walls of bare rock rise one above another, like the 
tiers of some cyclopean masonry, and are separated by usually short intervening 
slopes, which are sparsely clothed with grass and moss, and sprinkled with 
tumbled blocks and debris. The greatest elevations are reached in the two 
largest islands, Osteroe and Stromoe, Slattaretind in the former attaining an 
elevation of 2852 feet, and Skiellinge Field in the latter of 2502 feet. # Many 
other hills in these two islands are over 2000 feet in height, and some approach 
within 200 or 300 feet of the dominating point. Indeed, the average elevation 
of Osteroe and Stromoe can hardly be less, and is probably more than 1000 feet. 
The other islands are equally steep and mountainous, but in none do the hills 
seem to attain a greater elevation than 2000 feet. Thus Stoiatind in Waagoe 
is probably not over 2000 feet in height ; Kalsoe in the north-east is 1817 
English feet, Kunoe 2000 feet, and Naalsoe opposite Thorshavn 1276 feet. In 
Suderoe some of the hills are more than 1700 feet high — one of them, Kvanna- 
field, we found to be 539 metres = 1786 feet. The mean elevation of all the 
islands (exclusive of Stromoe and Osteroe) must exceed 800 feet, and is 
probably not less than 900 feet. 

The coasts are usually precipitous, many of the islands having only a very 
few places where a landing can be effected. Store Dimon, for example, 
possesses but one landing-place, and even that is accessible only in calm weather. 
The west coasts that face the open sea are as a rule the most precipitous — the 

* The height of Slattaretind is given in some Danish geographies which I consulted in the islands, 
as 2710 feet (Danish) = 2789 feet English; Eorchhammer makes it 2816 Erench feet; and another 
authority gives it at 882 metres = 2894 English feet. The height adopted in the text is that ohtained 
hy Mackenzie and Allan. There is a similar uncertainty as to the exact height of Skiellinge Field ; 
some Danish geographies and gazetteers giving it as 2350 feet = 2418 English feet. The height 
mentioned ahove is taken from the Danish Chart, which in Danish feet is 2431 feet or 2502 English 
feet. This corresponds with the height of 763 metres given by some writers. 



222 DJt JAMES GEIKIE ON 

finest mural cliffs occurring in Strom oe, between Westmannshavn Fiord and 
Stakken. These cliff's range in height between 900 and 2000 feet, and at 
Myling the nearly vertical walls of rock are even 2277 feet high. Osteroe and 
the north-east islands show sea-cliffs which exceed 1000 feet in height, 
and similar lofty cliffs occur in Waagoe, Sandoe, Suderoe, and all the other 
islets. 

3. Cliaracter of Valleys. — The best denned valleys are often comparatively 
broad in proportion to their length. Followed upwards from the head of a 
fiord, they rise sometimes with a gentle slope until in the distance of two or 
three miles they suddenly terminate in a broad amphitheatre-like cirque. In 
many cases, however, they ascend to the water-parting in successive broad 
steps or terraces (Plate XIII. figs. 2 and 3), — each terrace being cirque-shaped, 
and framed in by a wall of rock, the upper surface of which stretches back to 
form the next cirque-like terrace, and so on in succession until the series 
abruptly terminates at the base, it may be, of some precipitous mountain. 
Occasionally the col between two valleys is so low and level that it is with some 
difficulty that the actual water-parting can be fixed. Such is the case with 
Kolfaredal between Hoi and Leinum-mjavatn in Stromoe, where a well-defined 
hollow passes right through the hills, leading from the head of Kollefiord to 
the sea at Leinum. The height of the flattened col in this hollow is only 259 
feet, yet it is overlooked by hills that exceed 2000 feet in elevation. A similar 
long hollow crosses the same island between Saxen and Qvalvig. In Osteroe, 
likewise, a long low-level pass serves to connect the head of Fundingsfiord with 
that of Skaalefiord. In Sandoe and Suderoe, similar appearances may be 
noted. Besides these more or less well-defined valleys, the mountains 
frequently show isolated amphitheatric cirques. Sometimes these cirques open 
directly upon the sea, and may possibly represent the upper terminations of 
valleys, the lower portions of which have been removed by marine erosion. 
As examples, I may refer to the wide cirque (half mile in breadth and same in 
length) in Osteroe, which is cut across by the sea-cliffs between Andafiord and 
Fuglefiord, and to the narrower but equally well-marked cirque-valley (one 
mile long by about |th mile broad) which is abruptly truncated by the steep 
cliffs of Tiodnenaes in Suderoe. Other isolated cirques, much smaller than 
these, but yet sometimes of considerable size, may often be noted on the 
mountain-slopes at heights of several hundred feet or even yards above the 
bottoms of the valleys into which they discharge their drainage. Good 
examples occur amongst the high grounds above Saxen, where the small upper 
cirques must be 1200 or 1400 feet above the broad valley into which their 
waters tumble down a series of cliffs and precipices. 

4. Lakes and Streams. — There are numerous lakes in the islands, but they 
are mostly of small size. The two largest occur in Waagoe — one of them being 



THE GEOLOGY OE THE F^EROE ISLANDS. 223 

about 4 miles long, by half a mile in breadth. These, however, we did not 
visit. None of those we saw exceeded half a square mile in extent, and most 
were much smaller. At Sand in Sandoe, there is a narrow lake-like expanse 
of water about 1^ mile in length, which appears to be almost on the level of 
the sea, from which it is separated by low flats and sand dunes. But nearly 
all the lakes occupy rock-basins. 

Streams are very abundant, but none is of much importance. Owing to the 
excessive rainfall,"' however, they must occasionally discharge very considerable 
bodies of water, and as we shall see in the sequel they are potent agents of 
geological change. 

III. Geological Structure of the Islands. 

1. General Dip of the Strata. — The geological structure of the islands is very 
simple. The principal rocks are bedded basalts, with intercalated partings and 
layers of tuff, and in Myggenees and Suderoe of clay, shale, and coal. The 
prevalent dip of the strata in the northern islands is south-easterly, at a low 
angle, as was first pointed out by Mackenzie. In the north part of Osteroe, 
however, the dip is towards north-east, according to Forchhammer, whose 
observations we were able to confirm, and the same geologist states that the 
dip in Myggenges is easterly. In Suderoe, again, the strata incline uniformly to 
N.N.E. Nowhere is there any trace of a westerly inclination, and the steepest 
dips are found in Myggenaes, Waagoe, and Suderoe. In the former, Trevelyan 
saj s it is near 45°, while Forchhammer, who is probably nearest the truth, gives 
it as 10°. Judging from the view we had of the cliffs of Waagoe, the clip 
appeared to be 10° or 12° in the west of that island, but it decreased towards 
the east. In Suderoe the dip varies from 2° or 3° to 15° or thereabout. The 
south-easterly dip of the strata in the northern islands is certainly less than 
Mackenzie makes it, and cannot be greater than 2° or 3° on the average. 
Forchhammer is unquestionably correct in his view, that the rocks of Suderoe 
and Myggenass are the oldest of the series, and it will be most convenient 
therefore to treat of these rocks first. 

2. Contemporaneous or Bedded Basalt-rocks of Suderoe. — The basalt-rocks 
of Suderoe do not show much variety. The most common rock is a dark blue, 
almost black, and sometimes brown, fine-grained crypto-crystalline anamesite, 
which is usually scoriaceous and slaggy above and below, and not infrequently 

* On an average of several years, there are only forty days in the year on which it does not rain. 
The annual fall is 78 inches. It is a common belief out of Faeroe that the islands are shrouded in fogs 
during the greater part of the year. This, it seems, is a mistake. "We were told by Danes who had 
resided for some years in the islands, that they had not experienced more mist and fog than in Denmark ; 
and the meteorological records which are kept show that fogs occur on only thirty-nine days in the year. 
They are worst in the beginning of summer. The best time to visit the islands is from about the end 
of July to the end of August or middle of September. 



224 DR JAMES GEIKIE ON 

contains irregular layers of amygdaloidal cavities, ranged along the central or 
middle zone of the bed. It is composed of plagioclase, augite, magnetite, and 
olivine, the last being often more or less altered into serpentine. Such is the 
average character of the rocks which immediately overlie the coal-bearing 
strata, and anamesites of similar appearance predominate in the island. They 
generally decompose with a rusty brown crust, and are frequently much broken 
up, weathering into irregular spheroids. On slightly weathered faces, the rock 
has often a dull greenish colour and somewhat earthy appearance. The most 
distinctly amygdaloidal portions of a rock, are usually of a paler shade, and 
more close-grained texture than the darker less porous areas by which they 
are surrounded, and, viewed from a little distance, the various parts of one and 
the same flow resemble a series of separate beds. This peculiarity, which 
is sufficiently striking, was noted by Forchhammer, who remarks that "the 
amygdaloidal rock alternates with the basalt, but these alternations shade into 
each other, and are not at all well-defined, but are very distinctly seen when 
the rock is looked at from a distance of some hundred feet. Then the different 
layers are seen with their different colours, and one finds that the junction line 
is parallel to the plane which the principal mass of the dolerite forms with the 
clay stone (i.e., tuff), and therefore parallel to the stratification." 

The dark anamesite frequently becomes more coarsely crystalline, so as to 
assume the character of a typical dolerite, which is often rendered more or 
less porphyritic with plagioclase. This porphyritic character, however, is 
certainly much less common and less distinctly marked in the basalt-rocks of 
Suderoe than in those of the northern islands. Forchhammer, indeed, maintains 
that the traps above the coals are markedly porphyritic with " glassy felspar," 
while those underlying the coals are not porphyritic. Accordingly, he has in 
his map coloured all the northern islands and certain parts of Suderoe and 
Myggenees as " dolerite-porphyry" — the remaining portions of Myggenses 
and the southern island, which are below the horizon of the coal, being 
distinguished as " dolerite without glassy felspar." We could not, however, 
trace any difference in Suderoe between the basalt-rocks below and those 
immediately above the coals. At the same time, well-marked porphyritic 
and coarsely crystalline dolerites do occur in Suderoe at some distance 
above the coal. In the mountain called Nakin, for example, there is a 
highly crystalline and porphyritic dolerite quite like many which occur in 
Stromoe and Osterbe. No hard and fast line, however, like that suggested by 
Forchhammer, can be drawn between the beds above and those below the coal. 
The most that can be said is simply this, that while highly porphyritic 
dolerites prevail above the horizon of the coal, and therefore throughout the 
northern and smaller part of Suderoe, and over all the northern islands, dark 
non-porphyritic and fine-grained rocks are the most common varieties met 



THE GEOLOGY OF THE F^EROE ISLANDS. 225 

with below that horizon, so that anamesites predominate in Suderoe, and 
dolerites in the northern islands. 

The anamesites of Suderoe are, upon the whole, less strikingly amygda- 
loidal than the basalt-rocks of the northern islands, and the cavities seldom or 
never attain the large size that is frequently to be seen in the rocks of Stromoe 
and Osteroe. They are generally, but by no means always, drawn out in the 
plane of the bedding, and have thus often a flattened appearance ; frequently, 
however, they are almost circular, but more commonly still, perhaps, their 
shape is ragged and irregular. I have mentioned the fact that amygdaloidal 
areas often traverse the face of the rock in the plane of bedding, so as to form 
more or less well-defined lines. They also occasionally show a kind of 
curled, coiled, or involved arrangement, as if the rock had been rolled over 
upon itself while in a plastic or viscous state. Some of the largest amygda- 
loidal cavities we saw were in the rocks of Nakin and on the sea-coast at Waag, 
where they contain very beautiful zeolites. I noticed here stilbite, chabasite, 
quartz, and calcedony. Heulandite is said also to occur in amygdaloidal 
cavities in Suderoe, and chlorophseite is found at Qvalboe. Probably other 
minerals are to be met with, for I made no special search. According to 
Trevelyan, native copper is very frequent, though not abundant. Near 
Famarasund he found it in thin plates in a bed of "claystone ;" some of it, 
he says, contains gold, which is "also, but rarely, found separate." The place 
referred to by him is near the base of a sea-cliff which the boatmen pointed 
• out to us as we sailed past, but we could not stop to visit it. 

The upper and under surfaces of the anamesites form an interesting study. 
Sometimes the upper portion for several feet in thickness appears to be 
composed of a jumbled mass of irregular-shaped fragments of scoriaceous rock, 
which gradually shades, as it were, into the denser, non-porous crystalline 
mass of which it forms the crust. In other cases, however, the slaggy part 
appeared somewhat distinctly marked off from the denser rock on which it 
seemed to rest. Here and there the beds show a wrinkled and crumpled 
surface like that assumed by viscous bodies in motion. Some anamesites, 
again, appeared to have little or no scoriform crust, but were only somewhat 
amygdaloidal and vesicular atop, — the rock then having a tendency to weather 
into spheroidal masses. This latter character, however, was more frequently 
characteristic of the lower or under surfaces of the beds. These basal por- 
tions, so far as I had opportunity in Suderoe of observing them, appeared to 
be less scoriaceous than the upper surfaces, and they were often much less 
amygdaloidal and vesicular than the central part of the same flow. Sometimes 
one might detect bits of red tuff and shale caught up in the under surface of 
an anamesite, and very often the beds showed strong red discolorations 
below, when they came into contact with a pavement of tuff. Now and 

VOL. XXX. PART I. 2M 



226 DR JAMES GEIKIE ON 

again, too, the under part of a basalt-rock would present a highly broken and 
jumbled appearance — crystalline, compact, and non-amygdaloidal rock being 
commingled with highly vesicular and honeycombed fragments, but all welded 
together so as to form one solid mass. 

Many of the.anamesites are more or less distinctly columnar. Good examples 
of such are seen in the Trangjisvaag district, in the valley above Howe and 
along the sea-coasts. In Kvannafield these columnar rocks break up into 
fantastic walls and isolated peaks and tors not unlike ruined masonry. Even 
when true prismatic and columnar structure is wanting, the rocks are yet 
traversed by well-marked vertical joints, which, as will be pointed out 
more fully afterwards, greatly assist the denuding agents in their work of 
destruction. 

The beds appear to vary much individually in thickness, but I think 40 
to 70 feet would be a good average. Some, however, were certainly not 
over 30 feet thick, while others may have reached and even exceeded 100 
feet. We did not trace out any one particular bed to see how far it would 
extend, but could quite well follow the lines of bedding along the slopes of 
the hills, and could thus carry the outcrop of a particular anamesite for a 
distance of several miles. The rocks, however, have so much in common that 
I doubt whether in most cases one could surely identify the same bed in 
different valleys, unless the outcrop itself were actually followed. The aname- 
site overlying the coal-beds appeared to be one and the same flow, for 
wherever we examined it the rock showed a very similar character — the only 
differences being merely such as frequently can be traced in closely adjoining 
portions of one and the same rock-mass. And this is equally true of the 
anamesite upon which the coal-bearing strata repose. The separate flows, 
however, thicken and thin out, and although I nowhere in Suderoe chanced 
to come across the terminal portion of a sheet, yet I have no doubt that those 
beds which did not measure more than 12 or 15 feet in thickness were 
merely the attenuated terminal portions of much thicker flows. 

3. Bedded Tuffs. — The anamesites are usually separated from each other 
by tuff which varies in thickness from mere thin layers of a few inches up to 
beds 30 or 50 feet in thickness ; in some places the tuffs are even thicker. 
These tuffs are generally fine-grained, like indurated mud, but occasionally 
they pass into a kind of tufaceous grit. They are generally of a bright brick- 
red colour, but sometimes they are grey, blue, green, and yellow. In some 
good sections they are seen to consist of alternate ribbons and bands of fine- 
grained tuff, with shattery, crumbling, fissile, shaly clays. Often, however, a 
bed of tuff will show no lines of bedding, and looks like an amorphous 
structureless mudstone. Even in such beds, however, it will often be found 
that the tuff splits most readily along the plane of bedding, and a close inspec- 



THE GEOLOGY OF THE F^EROE ISLANDS. 227 

tion will sometimes reveal several lines of coarser granules alternating with 
the finer-grained sediment. Under the microscope, the red tuffs are seen to 
be made up essentially of so-called " palagonite." 

In the numerous exposures we visited I never was able to find any true 
lapilli. In a coarse-grained tuff that crops out on the eastern slope of Kvan- 
nafield a few small stones were detected, but none of these exceeded half-an- 
inch in diameter. Only in one place did we come upon what seemed to be a 
true agglomerate. On the shore at Qvalboe, the low cliffs, 10 to 12 feet 
high, are formed of an agglomeratic tuff resting upon anamesite. The matrix 
of comminuted gritty ddbris is crammed with angular and subangular stones 
and lapilli of ail shapes and sizes, from mere grit up to blocks more than one 
foot in diameter. The included fragments were all of basalt-rock, and the 
mass showed no trace of bedding. It is just possible, therefore, that it may 
be only the scoriform brecciated upper surface of an anamesite. 

4. Coal and Coal-bearing Beds. — The strata immediately associated with the 
coals of Sucleroe may be described as dark carbonaceous shales and clays, 
which frequently have a tufaceous aspect. But their general appearance will be 
gathered from the sections given in Plates XIII. and XIV. figs. 5-9, and described 
in the Explanation. The coal occurs as more or less lenticular layers in beds 
of dark indurated clay and shale. The seams, therefore, are very inconstant, 
and thicken and thin out in the most irregular manner. They appear mainly 
along one horizon, occupying a position about midway between the top and 
bottom of the trappean rocks of Sucleroe. This is the only horizon along which 
they have been worked to advantage, and it is doubtful whether they occur 
anywhere else in a workable condition. About 1100 feet lower down, however, 
another outcrop of coal occurs, but it appears to be very local. The section 
seen is shown in the illustration (Plate XIII. fig. 7). The whole thickness of 
the coal-bearing beds at this place was 10 to 15 feet. They consisted of fine- 
grained greenish mudstone and tufaceous shales, with some nodules of coarse 
ironstone. Forchhammer says that a layer of glance coal, 3 inches thick, was 
got here, but we saw cmly fragments of it lying about. This appears to be the 
lowest horizon at which coal has been met with. 

It is in the central and northern part of the island where coal has been 
principally worked, and there can be no doubt that all the workings are upon 
one and the same horizon (see Plate XIII. figs. 1 and 4). Trevelyan knew 
this, and the section which accompanies his paper gives a correct generalised 
view of the geological structure of Sucleroe. 

The coal is of two kinds : one a bright lustrous coal, which does not soil 
the fingers, having a glassy fracture, and closely resembling in general appear- 
ance some of the glossy parrot coals of the Scottish coal-fields — the other, a 
dull lustreless coal, which soils the fingers, and in which one may readily detect 



228 DR JAMES GEIKIE ON 

a vegetable structure. These two kinds of coal alternate in one and the same 
seam (Plate XIII. figs. 5 and 6) — sometimes a bed of glance coal being streaked 
with laminae of dull slaty coal, at other times a seam of slaty coal showing 
many thin lines of glance coal. This last, indeed, appears to be most usually the 
case, as the slaty coal is the commoner variety of the two. Johnstrup states 
that every lenticular mass of glance coal represents the flattened trunk of a 
tree, in which can be seen the annual rings of growth. This, I do not doubt, 
may be true of the thicker layers, but it does not seem to be the case with 
the finer streaks and threads ; at all events I could see no trace in these of 
flattened stems. But our examination was necessarily imperfect and in- 
complete. It seemed to me, however, that the alternating layers of bright 
and dull coal spoke to the gradual accumulation in water of different kinds of 
vegetable matter or of different parts of the same plants, and that the coal 
was analogous to what is sometimes seen in our Scottish coal-fields, where 
thin layers of gas-coal, black-band ironstone, and common coal alternate and 
interosculate in one and the same seam. 

The comparative composition of the two kinds of coal is shown in the 
following analyses, which are taken from Johnstrup's paper : — 

Glance Coal. 

I. II. 

Organic, 85-3 831 

Inorganic (Ash), ... 2-5 2 - 5 

Hygroscopic Water, . . . 12-2 14*4 

Common or Slate Coal. 

Good Coal. Bad Coal. 

I. II. I. II. 

Organic, . . . 78'0 73-4 65'0 60-6 

Inorganic (Ash), . 107 9-2 16-2 29-3 

Hygroscopic Water, . 11-3 17'4 18-8 101 

Johnstrup tells us that some of the coals are extremely ashy, containing as 
much as 51 and even 74 per cent, of inorganic material. This latter, however, 
is rather a carbonaceous clay than a coal. In open-air sections it is not 
difficult to trace the passage from coal into shale — an appearance which, taken 
in connection with the general aspect of the beds, is strongly suggestive of the 
aqueous formation of the coal-seams. I saw no trace of a true underclay, and 
nothing resembling rootlets. Indeed I was rather surprised at the apparent 
scarcity of plant-remains in the shales. Dark vegetable impressions were 
observed, but there was nothing among these which could have been 
specifically determined. The shales and clays associated with the coals 
generally contain thin lines and layers of glance coal and dull common coal, 
and now and then small nodules of coarse grey clay-ironstone make their 






THE GEOLOGY OF THE F^ROE ISLANDS. 229 

appearance. The mode of occurrence of these thin lines and streaks of coal in 
the shale seems clearly to indicate deposition in water of vegetable debris and 
muddy sediment. The shales and clays are generally dark dull grey, but 
sometimes they are rusty brown ; in close contact with the thicker coals they 
are usually very dark or black. In some places, as at Syd i Hauge (see Plate 
XIII. fig. 8), they have quite a tufaceous aspect, are dull green in colour, and 
do not differ from the green tufaceous shales which are sometimes met with 
between separate beds of anamesite. 

The outcrop of the coal-beds is shown upon the accompanying map, and 
does not differ much from that given by Forchhammer. Johnstrup's map only 
indicates the areas over which, according to his opinion, the coals extend. He 
has also left uncoloured those parts of the coal-beds that are at a lower level 
than the sea, and consequently considerable tracts in the north and north-east 
of Suderoe are ignored as coal-bearing. In the mountainous tracts of Borga- 
knappen and Kvannafield the coal-bearing strata seemed to us likewise to have 
a wider extension than Johnstrup's map allows. The line of bedding could be 
quite well followed from Kvannafield round to Borgaknappen, and the same 
beds occupy a considerable area in Tuanahelgafield. Of course, I do not 
maintain that workable coal will be found everywhere along the outcrop given 
upon the map. The seams, as I have said, thicken and thin out irregularly, 
and in no part of the coal-field probably will they be found to preserve for any 
distance an equable thickness or even to be continuously present. The line of 
outcrop simply indicates the geological horizon of the coal — the outcrop of 
the shales and clays in which the coals are found. 

The dip of the strata in Suderoe is uniformly N.N.E. Between Waag and 
Kvannafield the inclination probably does not average over 2° or 3°. It increases 
slightly north, and at Frodbbenypen (see Plate XIV. fig. 9), it is as much as 11° ; 
at Kvanhauge the rocks dip at 10° to 14°. Owing to the lowness of the dip, it 
will be seen that the coal-strata occur as isolated cappings on the crests of 
the high grounds in the middle of the island (see Plate XIII. figs. 1 and 4). 

5. Coal, <%£. of Myggences and Tindholm. — The only other islands in which 
coal occurs are Myggenses and Tindholm,"" in Sorwaagsfiord. We did not 
visit these, but I may mention what Forchhammer says about them. The coal 
of Myggenses, according to this authority, occurs as a thin layer, from i of an 
inch to 5 inches thick, embedded in a brownish clay or chinch. Associated 
with this crunch is a black schistose clay which now and then contains reed- 
like impressions, like those which occur in the black shales or schistose clays 
of Suderoe. The whole thickness of the coal-bearing beds is from 3 to 6 feet. 
They occur at an elevation of 1000 feet. 

* According to an old writer (Hbnschel), whose MSS. are in the Eoyal Danish Archives, coal is 
said to occur in Gaasholm and in Waagb'e. But it seems this is doubtful (Johnstrup). 



'2oU DR JAMES GEIKIE ON 

In Tindholm coal and clay are found irregularly associated with basalt. 
Forchhammer's description, which is not very clear, would lead one to suppose 
that the beds are much disturbed, interrupted, caught up, and enclosed in the 
basalt. He gives a section in which basalt is seen terminating abruptly 
against clay, and in the latter he says there are sporadic lumps of basalt. 
Perhaps these " lumps " may be only veins seen in cross-section.* Many of 
the basalt-dykes which intersect the sea-cliffs of the northern islands are 
accompanied by what seem to be irregular lumps of black basalt completely 
isolated and embedded in the surrounding dolerite and tuff— but these are 
clearly only the sectional faces of smaller veins proceeding from the main 
dyke (see Plates XIV. and XV. figs. 14 and 15). 

According to Forchhammer, the coal-beds of Tindholm are perhaps on a 
different horizon from those of Myggenses. 

6. Subsequent or Intrusive Basalts of Suderbe. — Basalt occurs intrusively in 
Suderoe both in the form of sheets and veins or dykes. Very fine exposures 
of a sheet of prismatic basalt 30 to 50 feet thick are seen on the shore at 
Frodboe, where the rock is fine-grained and of a dark blue colour. It shows 
a few small round vesicles, which are usually filled with calcspar. The columns 
of this basalt vary in diameter from a few inches up to 1 foot 6 inches and 2 
feet. They are often extremely regular, and sometimes show beautiful radiat- 
ing forms. Similar prismatic basalt occurs at Kvannabotnir. There and at 
Kvanhauge irregular dykes, veins, and sheets are intruded among the aname- 
sites, tuffs, and shales. All these dykes and sheets are evidently part and parcel 
of one and the same irruptive sheet, which may be traced in the cliffs round 
into Qvalboefiord as far as Tiodnenses, and a similar irregular mass occurs on 
relatively the same horizon on the opposite side of the fiord at Qvalboe, from 
which place it continues along the coast for some distance down the fiord. At 
one place to the north-east of Qvalboe the veins and dykes proceeding from 
this mass are beautifully displayed in the cliffs as they traverse a bright red 
rock, which is probably tuff (Plate XIV. fig. 10). Close to Qvalboe the basalt is 
quite prismatic. It is most probable that all these intrusive basalts belong to 
one and the same intrusive sheet. At Frodboe the rock occurs a little below the 
horizon of the coal-beds (see Plate XIII. fig. 1), while at Kvanhauge the latter 
are highly confused and disturbed by it ; and it continues at the same level all 
round the coast to Tiodnenses. On the opposite shore of the fiord it reappears 
at Qvalboe, where it seems to occupy the position of the coal — the coal never 
having been observed at this place. These phenomena are quite in keeping 
with what we know of the intrusive basalt-rocks of the Scottish coal-fields, 
which are peculiarly prone to occupy the position of coal-seams. Not un- 
usually our miners find some particular coal destroyed over a wide district by 

* Trevelyan says that the coal and clay of Tindholm are " apparently enclosed in the trap." 



THE GEOLOGY OF THE F^EROE ISLANDS. 231 

the intrusion of a sheet of dolerite. Occasionally such a sheet will leave its 
usual horizon, and, after rising through a considerable thickness of sandstone 
and shale, will push itself laterally into an upper seam, and continue along that 
line for some distance until it may ascend to a yet higher coal which it will 
" burn " in the same manner as the others. The coal-beds in the Carboniferous 
series of Scotland would thus appear to have been "lines of weakness." In 
like manner, the coal-bearing beds of Suderoe would seem to have yielded 
more readily to the assaults of the intrusive basalt than the harder and less 
easily divided anamesites with which they are associated. It is remarkable, at 
all events, that nowhere else in the island do such intrusive sheets occur. We 
could hardly have missed seeing them had they been present, for each cliff and 
mountain-slope is a magnificent geological diagram, in which every detail of 
structure is graphically exhibited. 

Dykes and veins of basalt, however, were noted in the cliffs between 
Famarasund and Famoye. These seemed to trend N.W. by N., nearly in the 
same direction as the coast. They sent out numerous small veins and threads 
in all directions, and appeared to be of the same character as the similar dykes 
which we observed in great abundance throughout the northern islands. 

7. Contemporaneous or Bedded Dolerites of the Northern Islands. — As I 
have already indicated, the basalt-rocks of the northern islands (Stromoe, 
Osteroe, Sancloe, &c.) are upon the whole more coarsely crystalline than those 
of Suderoe, and rather dolerites than anamesites. But, just as in Suderoe 
we occasionally come upon sheets of coarse dolerite interstratified with aname- 
sites, so in the northern islands anamesites are now and again encountered 
among dolerites. The prevailing colour of the dolerites of the northern islands 
is some shade of blue, but there are many black, grey, red, and purple varieties. 
All are composed essentially of augite, plagioclase, and magnetite, and most 
contain olivine, which is often abundant. Some of them are extremely coarse 
in texture— the crystals of plagioclase with which the rocks are invariably 
porphyritic, sometimes reaching half an inch or more in length,* and being often 
developed in great abundance, so much so indeed as occasionally almost to 
obscure the matrix in which they are embedded. It is this markedly porphy- 
ritic character which serves to distinguish the basalt-rock of the northern islands 
from those which in Suderoe underlie and immediately overlie the coals. 
Among the most beautiful porphyritic dolerites which we saw were those of 
Skaapen in Sandoe, of Hoyviig near Thorshavn, and of the hills at Storevatn 
near Leinum. The weathering of the augite, when the crystals are distinguish- 
able, often imparts a reddish-brown tint to the rock on slightly weathered 

* Forchhammer says they sometimes reach an inch in length. This able geologist's geognostic 
descriptions are in general singularly lucid, and he has given in small compass what seemed to us a per- 
fectly accurate account, so far as it goes, of the principal features presented by the dolerites of these islands. 



232 DR JAMES GEIKIE ON 

faces. The olivine occurs either as dark or pale green rounded and amorphous 
granules or imperfect crystals, but very frequently it is more or less altered 
into serpentine. So abundant is the magnetite that it often strongly affected 
the compass, and one had to be on one's guard while taking the direction of 
dips and glacial stride. Some of the rocks at Eide in Osteroe were particularly 
magnetic. The fracture of the dolerite varies of course according to the 
character of the rock ; sometimes it is smoothly conchoidal, but more generally 
it is somewhat irregular, more particularly in the case of the highly crystalline 
and more coarsely porphyritic varieties, some of which have quite a hackly 
fracture. 

I selected for microscopic examination a number of specimens which might 
be taken as fairly representative of the rocks of Suderoe and of the northern 
islands. Like the anamesites, the dolerites of the northern islands differ chiefly 
in texture and the varying proportion of their component minerals. In some 
cases, as at Skaapen, the base of the rock is quite compact like that of a basalt ; 
in many others it is crypto-crystalline, like that of the anamesites of Suderoe ; 
while in yet others the texture is coarse and granular, and between these 
varieties there is every gradation. Again, some of the dolerites are much more 
abundantly and coarsely porphyritic than others. Occasionally the dis- 
seminated crystals of plagioclase are small and few in number, while in some 
rocks, as already mentioned, they are large and thickly interlaced. The 
crystals of olivine also occasionally attain a large size. Owing partly to these 
differences and partly to structural peculiarities, the dolerites are variously 
acted upon by the weather, as will be afterwards pointed out more particularly. 
I found that as a rule the more highly amygdaloidal portions of a rock suc- 
cumbed most readily to the denuding forces — the denser or less amygdaloidal 
areas often projecting beyond these latter for several feet. 

Amygdaloidal and non-amygdaloidal areas frequently alternate in more or 
less regular bands which are parallel to the plane of bedding, and several of 
these alternations might be observed in one and the same bed, the line of 
separation between the bands appearing at the distance of a few yards to be 
often well-defined (see Plate XIV. fig. 11). The matrix of the amygdaloidal areas 
was frequently finer grained than the other parts of the rock ; it varied also in 
colour, and was often dull and earthy, becoming soft, wacke'-like, and crumbling. 
The non-amygdaloidal bands, on the other hand, were generally coarser grained, 
crystalline, and hard. The more crystalline parts of a dolerite might thus be 
black or blue, while the amygdaloidal portions were pale red or brown, grey, 
yellow, or dirty green. Small amygdaloidal cavities often occurred more or 
less numerously in the harder crystalline bands, along the line of junction with 
the wacke'-like areas ; but, so far as my observations went, they quickly dis- 
appeared at the distance of a few feet from the junction-line, although a few 



THE GEOLOGY OF THE F^ROE ISLANDS. 233 

might now and again be detected through the body of the rock ; and even 
occasionally a sporadic area, more or less vesicular and honeycombed, might 
be seen completely enclosed in crystalline non-amygdaloidal rock. Such 
alternating and variously coloured layers, although the individual zones are 
frequently of very irregular thickness, have yet all the appearance of true 
bedding when viewed from a little distance. 

The amygdaloidal cavities vary in size from mere points up to hollows more 
than 2 feet in diameter. Many of them are round, others are more or less 
flattened and drawn out in the plane of bedding, while yet others are quite 
irregular, and seem to have been formed by the confluence of several vesicles. 
In some places the cavities are very abundant — the rocks being literally honey- 
combed with them. When such is the case they are generally small — the larger 
cavities being perhaps most common when the rock is least porous. Some of 
the largest we saw were on the east coast of Skaalefiord. Very frequently the 
cavities occur in more or less continuous lines or layers, parallel to the plane 
of bedding (see Plate XIV. fig. 12), a feature which may be noted again and 
again in the sea-coast sections, particularly along the north-west shores of 
Stromoe. Now and again also may be observed that involved appearance of 
the cavities which has already been described as occasionally visible in the 
anamesites of Suderoe. The prevailing amygdaloidal minerals are chabasite, 
stilbite, mesotype, apophyllite, analcime, quartz, chalcedony, calcspar, and green 
earth. It is not uncommon to find two, three, or even four different zeolites 
in one and the same drusy cavity. 

I have said that more highly amygdaloidal parts often alternate with 
harder non-amygdaloidal zones in one and the same bed. This, however, is 
far from being always the case. Sometimes the vesicular areas appear to be 
as durable as the other portions of a rock, and do not differ from these either 
in colour or texture. Frequently the dolerites seem to be tolerably homogeneous 
throughout — there being no trace of that alternation of zones just referred to. 
The under and upper surfaces, however, wherever they came under my observa- 
tion, were always more or less vesicular, and often highly slaggy and scoriaceous. 
These scoriaceous portions are very striking in appearance. They appear to be 
made up of jumbled and broken fragments of highly vesicular dolerite — varying 
in diameter from a few inches up to several feet, enclosed in a less vesicular 
matrix of the same material. At other times the matrix appears to be 
amorphous, earthy-like, and highly discoloured, generally becoming bright red, 
and showing occasional yellow and parti-coloured areas. These red discoloured 
areas so closely resemble the tuffs upon which the dolerites repose that it 
is sometimes hard to say where dolerite ends and true tuff begins. When 
the rock is highly porphyritic, however, the presence of the large crystals of 
plagioclase in the reddened portions usually enables one to detect the line 

VOL. XXX. PART I. 2 N 



234 DR JAMES GEIKIE ON 

of junction between the two rocks, which is frequently very irregular. 
This discoloration is probably clue to the decomposition of the augite and 
olivine, and the production of sesquioxide of iron, and the resultant rock thus 
resembles the " palagonite " of penologists. It seemed to me, however, that 
in some cases the tuff over which the old lava flowed had been caught up 
and commingled with the latter, as I have frequently observed to be the case 
Avith the porphyrites of Carboniferous and Old Eed Sandstone tracts in 
Scotland, as in the Cheviots, the trappean hills of Ayrshire, the Ochils, the 
Sidlaws, and other Lowland ranges. In these Scottish areas the under portions 
of the porphyrites often contain quantities of baked sandstone and mudstone, 
which have evidently been caught up while in the condition of soft sand and 
mud, and become inter-coiled and involved with the once molten rock. Some 
of the dolerites of the islands now under review appeared to be much more 
scoriaceous above and below than others. Occasionally the slaggy under surface 
did not measure more than a foot or two, while in other cases it would reach 
as many yards. The upper surfaces were likewise often scoriaceous, but, owing 
perhaps to the comparative absence of red palagonitic matter, they were as a 
rule less conspicuous than the under surfaces. Some superficial crusts which 
I saw might readily have been mistaken for volcanic agglomerate. On the 
shore at Klaksvig, for example, a fine reddish brown vesicular dolerite is seen 
with a highly scoriaceous upper surface. This crust is made up of fragments 
chiefly vesicular, of all shapes and sizes, from mere grit up to pieces 6 inches 
and 1 foot in diameter. Some of the fragments were not unlike bombs, and 
had only the crust itself been visible it would have been difficult to distinguish 
the rock from a true volcanic breccia or agglomerate. Another appearance 
presented by the upper surface of some of the dolerites has been described by 
Sir G. Mackenzie as "not unlike coils of rope or crumpled cloth, an appearance 
which we should expect to be assumed by any viscid matter in motion." 

The beds of dolerite vary much in thickness, and it is not easy to give any 
average. Some were less than 20 feet thick, while others exceeded 100 feet. 
But as they do not preserve the same thickness throughout, it would only be 
possible to give a reliable average after a large number of individual beds had 
been followed along their entire extent, which has not yet been clone. It may 
be that the thinner beds seen in a section attain a greater thickness in some other 
part of their course, and that no single bed has a maximum thickness so incon- 
siderable as 20 feet. It is remarkable, however, for what a distance a bed will 
retain an uniform thickness. One could follow some well-marked sheets often 
along the whole course of a fiord, throughout which they seemed to retain very 
much the same thickness. As giving some notion of the number of beds, I 
may mention that in the hill-slopes and cliffs between Kollefiord and Kalbaks- 
fiord we counted some twenty sheets of dolerite, separated by lines and layers 



THE GEOLOGY OF THE F^ROE ISLANDS. 235 

of tuff — the highest bed visible being perhaps some 1500 feet or so above the 
sea-level. Again, in the fine mural cliffs and bare rugged slopes of Skiellinge 
Field, as viewed from Hoi at the head of Kollefiord, some thirty beds of basalt- 
rock and tuff-partings were visible, which would give an average of about 80 
feet for each bed. Some of the beds, however, were considerably under that 
thickness, while others could hardly have been less than 120 feet or even more. 
None of the bedded dolerites that we saw was so markedly columnar as the 
more typical of the anamesites of Suderoe. Now and again, however, we 
observed a rude approximation to columnar structure, and the rocks were very 
generally traversed by strong master-joints at right angles to the plane of 
bedding. 

8. Bedded Tuffs of the Northern Islands. — The tuffs of the northern islands 
are- precisely similar to those of Suderoe, and they need not therefore be 
specially described. ' I may merely state that we never chanced to come across 
any clunch or clay similar to that with which the coals of Suderoe are inter- 
bedded, and no trace of coal has ever been met with in the islands at present 
under review. The thickness of the tuff beds is very variable ; sometimes they 
consist of mere lines, while in other cases I saw in the cliffs beds which may 
have exceeded 100 feet in thickness. It is just possible, however, that these 
tuff-like beds, which were merely observed as we boated past, may have been 
earthy decomposing dolerites. Now and again I saw a tuff thin out as in the 
sea-cliffs near Gritenaes in Stromoe (see Plate XIV. fig. 13). But upon the 
whole the tuffs are less conspicuous than the dolerites, for while the latter 
form cliffs and steep faces, the latter usually give rise to slopes and ledges 
which are covered over with debris and vegetation. Some of these slopes 
indicated a thickness of 200 or even 300 feet of soft rock underneath, but 
whether this thickness was all tuff or partly tuff and rotten dolerite, I cannot 
say. In other places again, particularly in the sea-cliffs in the north-west of 
Stromoe, the tuff seems to occur as mere thin lines and partings. Some very 
instructive sections showing the rapid alternation of tuff and basalt-rock are 
seen along the southern shores of Fundingsfiord. Here beds of red tuff, varying 
from a foot or less up to several yards, are interbedded with irregularly weather- 
ing dolerite, and show well the undulating surface over which the old lavas 
occasionally flowed (Plate XIV. fig. 12). 

9. Subsequent or Intrusive Basalts of the Northern Islands. — Two sheets of 
intrusive basalt have been described as occurring — the one in Stromoe and the 
other in Osteroe ; but neither of these was visited by us. They are mentioned 
by every writer who has treated of the geognosy of the northern islands. The 
more striking of the two masses is that which is seen exposed along the 
western face of Skiellinge Field. It is a columnar basalt of an average thick- 
ness of 100 feet, which traverses the beds obliquely, and is represented by 



230 DR JAMES GEIKIE ON 

Trevelyan iii a section as extending from Norderdahl to Leinum. Its general 
features are well described by Allan. Another and thicker basaltic mass is 
figured by Trevelyan as occurring near Zellatrae (Selletrod) in Osteroe. 
These, so far as I kno"\v, are the only intrusive sheets which have been 
observed in the northern islands. Vertical or approximately vertical dykes 
and veins, however, are exceedingly numerous. We saw many in the cliffs, and 
these are indicated upon the map, and there are probably many more in those 
regions which we did not visit. We also came upon fragments of close-grained 
blue basalt in many places upon the hill slopes, which had doubtless been 
derived from dykes and veins exposed to denudation on the steep precipices 
above us. All the dykes seem to consist of the same kind of rock — a hard, 
fine-grained compact blue basalt (of the same composition as the bedded 
basalts), abundantly jointed at right angles to its direction, with several more 
or less continuous joints running parallel to its trend. The cross-joints give 
the rock quite a prismatic structure, the prisms being confined between the 
parallel joints, or between these and the walls of the dyke. Thus in one and 
the same dyke there may be several series of prisms ; but as both the dykes 
themselves and their parallel jointing are very irregular, the prisms are 
irregular also. Each dyke sends out numerous small veins which ramify in 
all directions, and are invariably minutely jointed across. The connection 
between these veins and the main dyke is often clearly exposed; but very 
frequently this is not the case, and the small threads and veins then appear 
quite isolated, the connection with the dyke having either been removed by 
denudation, or being still concealed behind the visible surface of dolerite 
through which the veins are squirted. Owing to their highly jointed character, 
these dykes fall rapidly before the action of the weather and the denuding 
agents. Thus in the sea-cliffs they almost invariably give origin to caves. 
Nothing can well be more striking than the appearance presented by many of 
these curious dykes. In the sea-cliffs they generally appear superficially black 
or dark green, and contrast very strongly with the ruddy coloured dolerites 
and tuff across which they break. I give sketches of a few of the more 
remarkable ones I saw (Plates XIV. and XV. figs. 14-17). Fig. 14 represents a 
set of dykes which are probably part and parcel of one and the same intrusive 
mass. All the appearances connected with this and other similar dykes gave 
one the impression that the basalt at the moment of intrusion must have been 
in a condition of extreme fluidity. Nowhere that we saw did the dykes disturb 
the bedding — they seemed to have cut the dolerites very much as a knife cuts 
cheese. Another very remarkable dyke is shown in fig. 16 a, b, Plate XV. 
The decomposition and erosion of this dyke have given rise to a curious cave 
which forms a kind of natural arcade with a double entrance, as shown upon 
the plan (fig. 16 a). The dyke is in the form of a cross, and the two limbs 



THE GEOLOGY OF THE F^ROE ISLANDS. 237 

meet in the roof of the arcade where the beknotted mass projects downwards, 
reminding one of the groined ceiling of some Gothic structure. But the 
eccentricities of these dykes are endless, and no two are ever alike. They 
strongly reminded me of the irregular basalt-dykes and veins which occur so 
numerously in the Outer Hebrides.* 

There appear to be two systems of dykes, but they probably belong to the 
same age. One series trends from a little east of north to west of south, and 
the other from north of west to south of east ; but the precise direction of 
some of those which we saw is not quite certain. When a dyke is only seen 
in one vertical sea-cliff section, and has not been traced inland, its true direc- 
tion may easily be mistaken, and such may quite well be the case with some 
which are indicated upon the map. Many of them, however, were so placed 
as to show their trend conspicuously enough, and these certainly gave evidence 
of a double series, one set running at approximately right angles to the other. 
But until the dykes in all the islands have been carefully mapped out this 
point will not be definitely settled. I may remark in passing, that they are 
never so regular as the Miocene dykes in Scotland; these latter, as is well 
known, traverse the country in great wall-like sheets, from which often few or 
no subsidiary veins proceed ; but the dykes of the Fseroes divide and subdivide 
again and again, and send out veins and threads innumerable. 

There is no evidence to show whether or not there is any connection 
between the dykes and the larger intrusive masses of prismatic basalt. It is 
quite possible, however, and even probable, that both belong to the same 
period of volcanic activity, and that they may have been injected from below 
at a time when sheets of basalt still continued to be poured out at the surface. 

IV. Thickness of the Strata : Conditions under which they were 

AMASSED. 

1. Thickness of the Strata. — The dip of the basalt-rocks and tuffs in the 
northern islands, exclusive of Myggenees, is somewhat persistently towards 
south-east, at an angle which hardly ever reaches 5°, and is usually much less. 
Sometimes, indeed, the clip is barely appreciable, and the beds appear to be 
quite horizontal. The average inclination certainly does not exceed 3°, but is 
probably greater than 2°. We have thus a thickness of 9000 or 10,000 feet 
for the rocks in the northern islands. To this must be added the thickness of 
the lower series (the beds under the coal) as developed in Myggenses, Suderoe, 
and Munken, which cannot be less than 4000 feet. 

2. Igneous Hocks o/Subaerial Origin. — Sir George Mackenzie was of opinion 
that the " traps " of the Fseroe Islands were the products of submarine volcanic 

* Quart. Journ. Geol. Soc, vol. xxxiv. p. 854. 



238 DR JAMES GEIKIE ON 

action, and this view has been generally accepted by geologists. Indeed, the 
FaToe Islands are sometimes referred to as "a typical example of an up- 
heaved submarine volcano." The greater lateral extension of the basalt-beds as 
compared to their thickness is supposed to indicate a flow under heavier pres- 
sure than that of the atmosphere alone. The phenomena presented by the old 
basaltic plateaux of Frerbe, Iceland, and other countries, are contrasted with the 
appearances which are known to characterise the eruptive products of Hecla, 
Etna, Vesuvius, and other modern subaerial volcanic cones, and since these latter 
rarely or never form such vast successions of parallel and widely extended 
sheets of lava, the older basalts I refer to are believed to have been spread out 
upon the bed of the sea. But if we put aside the fact of their greater 
horizontal dimensions, we find that the basalt-rocks of the Fseroe Islands 
present most of the features which are commonly supposed to be characteristic 
of subaerial lava-flows. They are generally vesicular, and often scoriform above 
and below ; they exhibit layers and lines of pores and larger cavities, often 
flattened out in the plane of bedding ; now and again their upper surfaces have 
that slaggy, wrinkled, and crumpled appearance which is so typical of certain 
modern lavas ; while their middle parts are not more vesicular than are the 
central portions of undoubtedly subaerial flows. The absence of fragmental 
accumulations, such as volcanic breccias and agglomerates, is really no proof of 
the submarine orign of the basalt-rocks of Faerbe and other similar trappean 
plateaux. Such coarse accumulations are generally distributed round the 
immediate neighbourhood of the orifice from which they are ejected. The 
basalt beds of the Fseroes may quite well have been ejected from one or more 
central orifices, in which case the absence of fragmental materials would only 
serve to indicate that the focus or foci of eruption must have been at a con- 
siderable distance from the present islands. Or it may be that the whole 
series of basalt-rocks are the products of vast fissure-eruptions. But if this be 
their origin I certainly met with no direct evidence in its favour. None of the 
numerous dykes which we saw could possibly be the feeders that supplied the 
bedded basalts. Most of the dykes died out upwards— often wedging out in 
the middle of a basalt-bed or tuff — and the rock of the dyke could always be 
clearly distinguished from that in which it terminated. The dykes, in short, 
are only thin, irregular veins that ramify and split up into mere threads, and 
have no resemblance to those great wall-like basalt-dykes of supposed Miocene 
age, which are so common in Scotland. Veins exactly similar to those of the 
Fseroes, however, are very common in the Outer Hebrides. 

But whatever the particular origin of the bedded basalts of the Fserbe Islands 
may have been — whether they flowed from one or more foci like the lavas of 
modern volcanoes, or welled up from below along the lines of great fissures — all 
the evidence is against the view that they were erupted upon the bed of the 



THE GEOLOGY OF THE F^EROE ISLANDS. 239 

sea. If this had been their origin, we should be at a loss to account for the 
total absence of marine organic remains in the interstratified tuffs. Nothing 
at all resembling the fossiliferous tuffs of the Campagna di Roma and the Terra 
di Lavoro is to be found. Instead of these we have the coal-beds of Suderoe; 
and hitherto the only fossils which the Miocene volcanic rocks of northern 
regions have yielded are land-plants, which would be inexplicable enough if 
these igneous masses had invariably been the products of submarine volcanoes.* 
The equable thickness and wide extension of the bedded basalts, which have 
been thought by some to indicate that the old lavas have been spread out 
under the weight of a superincumbent ocean, are equalled and even surpassed 
by the great lava-flow from Skaptur Jokul in 1783, which covers an area as 
extensive as that of all the Fseroes, and which in the open country does not 
average more than 100 feet in thickness. It may be admitted that the lavas 
of some modern volcanoes have " a more rugged and porous aspect " than 
many of the basalt-sheets of the Freroes — amongst which we look in vain for 
those great " clinker-fields " and cinder-like masses which are often met with 
in the products of recent eruptions. But all modern lavas are not equally 
sooriaceous, and many have no thicker cinder-like crust than the old basalts of 
the region under review. 

While not disputing that the Freroe basalt-beds may have been poured out 
from fissures, it seems to me that the phenomena are not inexplicable on the 
view that they have proceeded from one or more foci in the manner of modern 
lavas. But if this has been the case, then it is obvious that the centre or 
centres of eruption must have been far removed from the site of the present 
islands. This, as I have already remarked, would explain the absence of 
breccias and agglomerates, of lapilli and bombs. It would also account for the 
uniform character of the ancient lavas. It is well known that in modern vol- 
canoes lavas and tuffs of very different character issue at different times from 
one and the same orifice, or from craters which are contiguous. Thus a 
volcanic mountain may be built up of a great succession of basalt-rocks, 
trachytes, trachy-clolerites, obsidian, agglomerate, fine tuff, &c. And the 
basalts of modern formation do not differ essentially from those of older tertiary, 
secondary or primary age — such differences as do occur being sufficiently 
accounted for by long- continued chemical action. It is likewise true that 
trachytic rocks are not confined to modern volcanoes, but occur also inter- 
stratified with tertiary and secondary strata. We may reasonably infer, then, 
that if the Fseroe basalts came from volcanic foci, and consolidated in proxi- 
mity to these, they ought to have been associated not only with coarse agglo- 

* Marine fossils are said to occur in the Surtarbrandr or lignite-beds of the sea-coast, near Husavik 
in Iceland ; but this appears to be exceptional — the palagonite-tuffs of that island being otherwise as- 
destitute of any trace of marine life as those of the Easroes. 



240 DR JAMES GEIKIE ON 

merates but with other varieties of lava. But as this is not the case, it may be 
supposed that the whole vast series of basalt-beds and tuffs (13,000 feet or 
14,000 feet in thickness) accumulated upon the outskirts of an old volcanic 
area. They would, in this view, represent the heavier and more fluid lavas, 
derived from foci which may also have ejected agglomerates and many lavas of 
lighter specific gravity — these last having been unable to reach the great 
distances attained by the basalts. This would only be upon a larger scale than 
what we know has taken place in regions like Auvergne, where, as " in the Mont 
Dore, the trachytic currents," according to Scrope, " have in no instance flowed 
more than from 4 or 5 miles from the central heights of the volcanoes ; 
the basaltic currents, on the contrary, have reached a distance of 15 miles or 
more." 

3. Miocene Age of the Strata : Physical Conditions under which they were 
amassed. — Although, as I have said, the plant-remains of Suderoe have not been 
specifically determined, there is no reason to doubt that geologists are right in 
referring the igneous rocks of the Faeroe Islands to the Miocene period. They 
almost certainly belong to the same great series of which the basalt-plateaux of 
Iceland, Greenland, Spitzbergen, and our own islands form separate portions. 
Such being the case, it may be allowable to offer a few remarks on the physical 
conditions under which the rocks of the Fseroe Islands would seem to have 
been accumulated. 

We know that during the Miocene period there existed a very wide extent 
of land in northern regions. It is even highly probable that America and 
Europe were at that time connected, so that plants could migrate freely across 
broad areas which now lie drowned beneath the waters of the Arctic Ocean. 
It is not unlikely that during Miocene times land may have stretched con- 
tinuously westward from what are now the Fserbe Islands to Iceland and 
Greenland. This belt of land must have been the scene of great volcanic 
activity, and we may conceive how after many successive sheets of lava had 
been poured out from one or more vents, or from long fissures, all the hollows 
of the old land-surface would be filled up for as great a distance as the molten 
rock flowed. If the lavas flowed from orifices like those of ordinary volcanoes, 
there may have been one or more central cones, rising probably to a consider- 
able elevation, and surrounded by vast plains that sloped outwards with a 
diminishing inclination in all directions. The cones themselves would be 
built up of irregular masses of different kinds of lava and heaps of more or less 
loose scorise, lapilli, bombs, and tuff. The same materials would also enter 
largely into the composition of the immediately adjacent low grounds. But 
the further one travelled from the centre of dispersion, the less abundant would 
lapilli and other loose ejecta become. Highly porous and scoriaceous lavas 
and clinkers would in like manner abound in the vicinity of the volcanic centre, 



THE GEOLOGY OF THE F^ROE ISLANDS. 241 

but they would become less conspicuous as the outskirts of the igneous area 
were approached. The lavas would still continue to show scoriform crusts, but 
many of them would begin to exhibit a somewhat smoother or less rugged 
surface, showing, in place of great fields of cinder heaps, a wrinkled and 
crumpled appearance, such as is assumed by any viscid substance in motion. 
In short, the outlying parts of the igneous region would be invaded only by 
the more fluid and specifically heavier lavas, the lighter and more porous lavas 
and agglomerates would in great measure be restricted to the cone or cones 
and their vicinity. Tufaceous deposits, however, would not be wanting in any 
part of the land to which the lavas might flow, and they might well extend 
even much further. The tuffs of the outlying regions, however, would generally 
be fine-grained, consisting of the volcanic dust which the winds had power to 
carry to considerable distances ; of volcanic mud or tuff ; and of the materials 
derived from the subaerial disintegration of the exposed lavas. Such fine- 
grained detritus, whether washed down the gentler slopes by rain or swept 
forward as mud, would tend to accumulate in the hollows of the ground. More- 
over, since the inclination of the surface in the outer zone of the volcanic area 
must have been as a rule very gentle, there would be no rapid flow of water, and 
therefore comparatively little aqueous erosion ; and thus coarse gravel and 
shingle would be generally absent. But, just because the slope of the land was 
gentle and tolerably equable, fine tufaceous alluvia would tend to be widely dis- 
tributed — gathering thickly in the hollows, and thinning off where the ground 
rose in swells and undulations. In short, they would form rather sporadic 
patches and layers of variable depth, than widespread continuous sheets of 
equable thickness. These assumptions, as we have seen, are confirmed by 
direct observation ; the tuffs are much less continuous and of less uniform 
thickness than the basalt- rocks with which they are interstratifiecl. 

Now let us suppose that, after many sheets of lava had been poured out 
and spread above the site of the future Suderoe, a pause in the volcanic activity 
ensued, and the region referred to ceased for a time to be overflowed. We can 
readily believe that, under the climatic conditions prevailing in Miocene times, 
vegetation would sooner or later creep over the surface of the cooled basalts. 
Rain-water would gather in the depressions of the ground, and so give rise to 
swamps and pools and shallow straggling lakes, in which mineral and vegetable 
matter would gradually accumulate. It is not improbable that the basalt- 
plateaux might even be densely wooded. Such conditions might obtain for a 
prolonged period, so as to allow of the accumulation in swamps, bogs, and 
lakes of very considerable depths of vegetable matter, mixed with mud and 
clay and fine sand, and now and again with small stones or pebbly grit. At 
other times the suspension of volcanic activity might not be so prolonged, and 
a renewed incursion of lava might take place before the last current had 

VOL. XXX: PART I. 2 



242 DR JAMES GEIKIE ON 

sufficiently cooled to permit of vegetation spreading over its surface, or indeed 
before the plants themselves had time to occupy the ground. Again, it might 
occasionally happen that, owing to the nature of the volcanic surface, there could 
be little or no accumulation of vegetable matter in swamps and pools, and the 
mere " carpet of greenery " which may have covered the ground in whole or in 
part, might well be destroyed upon the advance of another lava-flow. 

The appearances which have just been described as the most likely to occur 
under the circumstances I have supposed, are precisely those which are found 
associated with the coal-beds of Suderoe. The character of the clays and 
tufaceous shales with which these coals are interbedded, and the manner in 
which mineral and vegetable layers interosculate, all point to quiet deposition 
in shallow lakes, and the slow accumulation of plant-remains in swamps, 
marshes, and bogs. The thin layer of coal at Dalbofos may indicate a com- 
paratively short period of rest, while the thicker coals at the higher level seem 
to bespeak a pause of much longer duration. In like manner, many of the 
tufaceous shales which appear to be uufossiliferous, may yet have been accumu- 
lated in precisely the same manner as the shales which accompany the coal 
layers, and which in some places they closely resemble. The mere absence 
from these shales of coal or plant-remains does not necessarily prove that no 
vegetation covered the plateaux at the time the shales in question were accumu- 
lated. The presence of the little patch of coal at Dalbofos, which is quite 
local — none ever having been seen elsewhere on the same horizon — shows us 
how easily all trace of a vegetable layer might be obliterated. But while it 
may be true that some of the darker tufaceous shales may thus be of the nature 
of " surface-wash," and owe their origin directly to the action of the weather, 
there can be little doubt that by much the greater proportion of the red and 
particoloured tuffs are due more or less directly to igneous action. They 
consist of the finer dust and grit blown out during eruptions, and spread by 
the winds over vast areas, and partly no doubt of the same material carried down 
by rain and swept from higher to lower levels by running water — sometimes, 
perhaps, the tuffs may represent former currents and streams of volcanic mud. 

4. Position of old Volcanic Centre of Eruption. — It would be interesting to 
ascertain the locality of the volcanic centre from which the old lavas of the 
Fseroe Islands were ejected. I am inclined to think that it lay somewhere to 
the westward, partly for the reason that the rocks have an average easterly 
dip, and partly because the sea between the Freroe Islands and Iceland is not 
so deep as it is in the direction of Scotland. The dip, although it may have 
experienced subsequent modification, may yet indicate the original inclination 
of the ground, while the lesser depth of the sea to the west does not imply 
such extreme depression and denudation as must have taken place if the site 
of the ancient volcano lay far to the east of the present islands. At the same 



THE GEOLOGY OF THE E^EROE ISLANDS. 243 

time, there is something to be said for the view that the old volcanic centre may 
after all have occurred in that direction, and that owing to the sinking down 
of the central region, the dip of the great basalt plateaux has been reversed. 
Such enormous denudation and so many great movements of the earth's crust 
have taken place since Miocene times, that too much stress may easily be laid 
upon the configuration of the sea-bottom, and the present dip of the strata. In 
whatever direction the centre of eruption lay the fact remains, that in the basalt- 
masses of Faeroe we see only a few shreds of what must at one time have been 
a continuous plateau, occupying a much wider superficial area. The lofty 
cone or cones, if from such rather than fissures the basalts were erupted, have 
entirely disappeared — perhaps the looser and less consolidated materials of 
which they would probably be composed, having, we may suppose, contributed 
to their overthrow. This, I may remark, is the fate which has overtaken all 
the great volcanic centres of our own islands. Thus the volcanic products of 
the Old Red Sandstone, Carboniferous, and Permian formations are now 
represented chiefly by sheets of igneous rock, many of which have consolidated 
at a greater or less distance from the points of eruption. Such cones of scoriae 
and lava as may formerly have existed have been entirely demolished — 
their sites being now indicated by the hard plugs of rock that occupy the 
pipes or necks through which ashes and molten rock found a passage to the 
surface. 

V. Glacial Phenomena of the Islands. 

1. Early Notices of Glacial Phenomena. — The earliest notice of the abraded 
and striated rocks of the Fseroe Islands occurs as already mentioned in Mr 
Allan's paper. His description does not tell us in what direction the abrading 
agency moved ; but Robert Chambers, who visited the islands forty-three years 
afterwards, and saw the roches moutonnees and stria; described by Allan, 
was clearly of opinion that the ice had come from the north. He says, " They 
(the striae) presented themselves in abundance in several places, most strikingly 
of all within sea-mark on the shore of the quiet bay, being all directed from 
the north," &c. Again he describes similar striae observed by him in Iceland, 
which had the same trend with those at Eide, and concludes that these facts, 
taken in connection with observations of a like kind in the north of Europe 
and America, indicate " that there has been one universal sweeping of the 
surface by ice, down to some point in latitude which remains to be deter- 
mined. The parallel channels between the Faeroe Islands, all lying between 
north-west and south-east, I regard as excavations made by this wide-spread- 
ing arctic ice-sheet." Mr Chambers likewise noticed the glaciated rocks at 
Thorshavn, but failed to see the striae which many of them present. Again, 



244 DR JAMES GEIKIE ON 

lie chronicles the occurrence of striae at Westmannshavn, " directed from 
N. 80° E. (when 30° were allowed for variation)." Since Chambers' visit no 
geologist, so far as I know, has published any account of the glaciation of 
the islands. Professor Johnstrijp's interesting paper makes no reference to 
the " superficial geology " of Suderoe, and this paper, as I have said, is the 
latest contribution to our knowledge of the geology of the Faeroe Islands. 

I shall describe the glacial phenomena under the following heads : — 
Glaciation ; Till or Boulder Clay ; Erratics and Morainic Debris ; and Lake- 
Basins. Some further remarks on the subject will also be given when I come 
to discuss the origin of the valleys and fiords. 

2. Glaciation of the Islands. — Every island visited by us showed conspicuous 
marks of glacial abrasion. Sometimes the striae were finely preserved, at 
other times they were faint, and only the deeper ruts were conspicuous upon 
the smoothed faces, while in very many cases all the more delicate ice-mark- 
ings had disappeared, and only the characteristic rounded and dome-shaped 
outlines remained. Frequently, too, the roches moutonnees have been broken 
and shattered by the action of frost to such an extent that the glacial form is 
best seen from a little distance. Some of the most perfect examples of striated 
rock-surfaces occur in Stromoe. Thus at Thorshavn we detected a number 
of well-smoothed faces, the direction of the striae varying from E. 40° S. to 
E. 45° S., the abrading agent having clearly come from north-west. Some of 
these striated faces occur in the little town itself; as, for example, upon 
rounded rocks on the side of a street or road not far from the church. They 
are visible also upon roches moutonnees at various places along the shores of 
the little promontory, upon which a considerable part of the town is built. 
But perhaps the best example is seen upon the side of the path that leads to 
the fort. Here, at the distance of 90 or 100 yards from the latter, there is 
a wide surface of basalt planed down to a level and traversed by long parallel 
striae and ruts which trend E. 35°-45° S. In the immediate outskirts of Thors- 
havn several other striated faces were observed showing a similar direction ; 
but on one face the direction was more easterly — E. 10° S. Some good 
examples were also noted in the neighbourhood of Hoyviig, a mile or two 
to the north of Thorshavn. At this place fine roches moutonnees may be seen, 
and well-marked striae pointing E. 40° S., which occur both on horizontal, 
sloping, and vertical surfaces. The hills immediately to the west of the town 
show marks of glacial abrasion up to and over their summits (1048 feet), and 
the high ground between Arge and Kirkeboe appear in like manner to have 
been smothered in ice, the basalt-cliffs and terraces have been bevelled and 
rounded off, and the hill-slopes generally show a well-marked glaciated contour 
or outline. 

All the steep seaward slopes between Thorshavn and Kalbaksfiord are 



THE GEOLOGY OF THE F^ROE ISLANDS. 245 

highly glaciated, fine roches moutonnees being seen on the shore at Qvitenaes, 
where the tops of the columns of basalt are finely smoothed off. Although 
we only saw these rocks from the boat as we passed along the shore, we were 
yetnear enough to distinguish the coarser striae, which appeared to have the 
same direction as those at Hoyviig and Thorshavn. The appearance of the 
glaciation on the hill-slopes in Kalbaksfiord is very impressive. Here one 
can see at a glance in which direction the ice has flowed ; it has clearly 
crossed the lower reaches of the fiord from north-west to south-east, a direc- 
tion which corresponds with the trend of the upper part of the fiord. On 
the south side (the Stoss-seite) the hill-face exhibits the strongest marks of 
glaciation, while on the north side (the Lee-seite) the dolerites are rough and 
rugged, and show little or no trace of abrasion. The seaward slopes between 
Kalbaksfiord and Kollefiord also exhibit marks of glacial abrasion. 

We crossed Stromoe from Oreringe to Westmannshavn. The lower parts 
of the mountains that overlook Kolfaredal are smoothed and abraded in a 
south-east direction, and we estimated the height reached by the glaciated 
outline to be some 1500 or 1600 feet. Above that level all is rough and 
rugged, and destitute of the slightest trace of glacial abrasion. At the 
Storevatn of Leinum we found the roches moutonnees at the exit of the lake 
gave evidence of an ice -flow towards the south-west into Westmannshavnfiord. 
The pass across the dividing ridge between Kolfaredal and the valley that 
leads down to Westmannshavn we found to be 1243 feet (379 metres). At 
this level are roches moutonnees, but we saw no striae. The glaciated outline 
was continued up the mountain-slopes above us for not less than 400 feet. 

At Westmannshavn many well glaciated surfaces occur, but the striae have 
in most cases disappeared. In one or two places, however, upon the steep 
hill-slopes to the west of the large waterfall, faint striae and ruts were observed 
with a trend of W. 30° S., while close to the waterfall itself we got them 
pointing S. 5° W. Again, upon a point that juts into the sea E.S.E. from the 
church, ruts and striae, directed to S. 40° W. and S.W., occur upon the surface 
of roches moutonnees. All the hill-slopes surrounding the bay are highly 
abraded, the basalt-cliffs and terraces being rounded and smoothed off in a 
striking manner (Plate XV. fig. 18). 

The valley that opens upon Saxenfiord is likewise well glaciated, and 
exhibits smoothed and striated rocks in several places. On the plateau near 
the church many smoothed surfaces appear, but the striae have in most cases 
vanished. We got several good examples, however, all of which pointed in 
the same direction, namely, W. 25° N. or down the valley. Between the 
church and the lake we met with other instances, but the roches moutonnees 
were, as a rule, much broken up, and to a large extent masked by their own 
ruins. We traced the glacial outline in the district between Saxen and 



246 DR JAMES GEIKIE ON 

Tiornevig up to within 100 feet or so of the col or water-parting, immediately 
above which the rocks give no evidence of having been subjected to glacial 
abrasiou. The col we found to be 1693 feet (516 metres) above the sea, and 
the glaciation came close up to this level ; the rocks upon the col, however, 
were much broken up by frost, but abraded rocks with the characteristic 
glaciated contour certainly reached 1600 feet. We saw no striae at Tiornevig, 
but the sea-ward slopes of Stromoe opposite Osteroe show well-marked 
roches moutonnees between Tiornevig and Haldervig, and much further south, 
as we could see very plainly from the high grounds of Osteroe. 

Perhaps the best preserved roches moutonnees we anywhere observed were 
in Osteroe and Sandoe. It was with considerable interest that we visited the 
northern portion of the former island, for we felt that the evidence to be 
gathered there would go a long way to settle the question which we had come 
to solve. No difficulty was experienced in finding the locality described so 
long ago by Allan, and subsequently visited by Chambers, but the striae, 
instead of being " directed from the north," had clearly been graved by ice 
coming from quite the opposite point of the compass. The Kodlen peninsula 
we found glaciated all over, the roches moutonnees on both sides of the isthmus 
being beautifully perfect, and showing Stoss- and Lee-seiten in the most 
admirable manner. In many places the striae are well seen, and long ruts and 
channelings or grooves and trenches, well smoothed and ice- worn, traverse the 
rock-surface. The direction of the striae, ruts, and grooves varied a little from 
N. 10° W. to N. 10° E. — the variation being evidently due to the form of the 
ground. 

We traced the glaciated contour up to a height of 1302 feet (397 metres), 
which was the summit level of the pass leading from Eide to Funding, but the 
slopes facing the sound between Osteroe and Stromoe seemed to be glaciated 
to a somewhat greater height. The direction of glaciation upon those slopes, 
so far as we could observe them, seemed to be in a direction corresponding 
with the trend of the sound, namely from S.S.E. to N.N.W. Crossing the 
ridge to Funding, we found that the glaciation pointed east into Fundingsfiord, 
and that ice had evidently gone down the valley. The rugged mountains 
overlooking the upper reaches of Fundingsfiord from the east appeared con- 
spicuously glaciated in the direction of the fiord, but the upper parts of the 
rough hills between that and Andafiord were above the limits of glaciation. 
At Andafiord we got striae upon a surface of basalt under boulder-clay. The 
striae pointed down the fiord or E. 40° N. 

The rugged promontory between Leervigsfiord and Giotheviig shows strong 
marks of glacial abrasion in the direction of those fiords, but the higher parts 
of the ridge project above the glaciated area. The southern shores of Giothe- 
viig are well rubbed in the same direction. Between Giothe and Skaalefiord 



THE GEOLOGY OF THE F^EROE ISLANDS. 247 

the roches moutonnees are well defined, and show striae, ruts, and grooves, 
which point E. 35°-40° N., evidently the work of ice which overflowed from 
Skaalefiord. The dividing-ridge between Skaalefiord and Giotheviig, at the 
place we crossed, was 410 feet high, but the glaciation swept up to within a 
short distance of the hill-tops. Skaalefiord itself has brimmed with glacier-ice, 
the great body of which flowed down the fiord, as the highly abraded seaward 
slopes on both sides clearly attest. The glaciation is particularly well seen at 
Tofte Naes — the whole of that peninsula exhibiting every evidence of severe 
glaciation. The direction of ice-flow, as shown by the roches moutonnees, was 
towards S.S.E. The west coast of Osteroe, opposite Kollefiord, also gives 
evidence of having been abraded in a south-east direction. 

I have no doubt that glacial stria? might be found in other parts of Stromoe 
and Osteroe which we did not visit. In the neighbourhood of Qvalvig, for 
example, glacial phenomena are probably well developed. But the localities 
we examined sufficed for our purpose, and supplied abundant evidence to show 
that these islands had been glaciated by a local ice-sheet. We found not the 
slightest indication that they had ever been impinged upon by ice flowing from 
the north. 

The evidence obtained in the smaller islands served further to establish this 
conclusion. We visited Boroe and found that Boroevigfiord was abraded in a 
S.S.E. direction — in other words, the ice had flowed down the fiord. We 
boated along a portion of the coast-line of Kalsoe and Kunoe, but the other 
islands, Wideroe, Svinoe, and Fugloe, we did not approach. We could see, 
however, that the higher parts were extremely rugged and quite destitute of 
any appearance of glaciation ; and from the analogy supplied by Boroe, I have 
no doubt that their lower portions will give evidence of a local ice-flow. 

Naalsoe, opposite Thorshavn, appears smoothed off to its summit ; and 
the seaward slopes of Waagoe opposite Westmannshavn are glaciated in 
the direction of Westmannshavnfiord. But I could not be sure whether 
the ice in this fiord had moved to north-west or south-east. It is most 
probable, however, that the ice-flow was in both directions — in the northern 
reach going towards north-west, and in the southern section to east and south- 
east. The rugged mountain of Stoiatind in Waagoe soars above the limit of 
glaciation. 

The direction of glaciation in Hestoe we did not determine, but it will doubt- 
less be found to agree with that of the hill-slopes on the Stromoe side of Hestoe- 
fiord, which, so far as I could make out, was towards S.S.E. Unfortunately, 
thick fog prevailed when we traversed the district between Thorshavn and 
Kirkeboe, and we were not fortunate enough to find any stria? at the latter 
place. 

At Skaapen, in the north of Sandoe, the ground is highly glaciated, but 



248 DR JAMES GEIKIE ON 

owing to the absence of any well-defined Lee-seite, and the disappearance of the 
strite, we could not be certain as to the direction followed by the ice. On the 
opposite side of the island, however, we found strongly marked roches 
moutonnees, and very fine examples of striation. As these are perhaps the best 
preserved specimens to be found in the Faeroe Islands, it may be well to 
indicate the precise locality. We met with them at the point which forms the 
south-west limits of Sandsbugt. Close to this point there is a deep ragged cleft 
in the rocks into which the sea has access by a subterranean passage. The 
dolerite at this place shows fine striae pointing to S. 40° W., but the best example 
occurs on the headland at the point itself. Here the roches moutonnees indicate 
very clearly the direction of the ice-flow, and the striae (S. 40° W.) are particularly 
sharp and fresh. Nearer the village of Sand, we found striae with a more 
southerly trend — S. 15° W. Of the interior part of Sandoe I can say very little 
— for we traversed it in a dense drizzling fog. We could only see that roches 
moutonnees and ice-worn rocks accompanied us across the hills. 

The higher parts of Skuoe, as seen from our boat, appeared to be smoothed 
off from the north or north-east ; but Store Dimon and Lille Dimon, when we 
passed them the first time, were shrouded in mist, and on our return from 
Suderoe rough weather prevented us approaching them. 

Suderoe has supported a considerable mass of ice, for we traced the 
glaciated outline up to a height of 1400 feet. Above that level all is rough, 
angular, and serrated. The low ground that extends from the head of Qvalboe- 
fiord to the west coast is highly moutonn^e, the position of the Stoss- and Lee- 
seiten indicating an ice-flow from east to west. Here also the striae point 
E. and W. In Trangjisvaag valley, the direction of glaciation is towards south- 
east, as shown both by roches moutonnees and striae. Both sides of the fiord 
into which this valley opens are highly glaciated in the same direction. At 
Ordevig, the striae point E. 30° N., and correspond in direction with the trend of 
the valley in which they occur. The fine cirque-like valley of Howe affords 
admirable examples of glaciation. The whole broad amphitheatric space has 
been filled with ice, like a great reservoir ; the flat bottom being thickly set 
with roches moutonnees, and the smoothed and rounded glacial contour rising 
on the hill-slopes to a height of 1400 feet. The upper part of the valley is 
sprinkled with many lakelets, which rest in true rock-basins. Striae are not 
abundant, but we noticed them in several places, and they all pointed to the 
east, or down the valley. Another finely glaciated cirque valley descends from 
Kvannafield and Borgaknappen to the cliffs on the west coast, north of Fama- 
rasund. The ice that filled Howe valley must have brimmed over and become 
confluent, not only with the Trangjisvaag ice, but also with the glacier masses 
that descended the Dalbofos valley and Waagsfiord, for the hills above Porkerji 
and Naes are strongly glaciated all over. The trend of the abraded rocks on 






THE GEOLOGY OF THE F^EROE ISLANDS. 249 

both sides of Waagsfiord is towards the south-east, but at Waag there is a hollow 
which runs from the head of the fiord south-west to the open sea coast, along 
which a stream of ice has flowed, as is shown by roches moutonnees and striae 
pointing S.W. At Famoye, likewise, we got evidence of an ice-flow to the west 
— striae and roches moutonnees on the south side of the bay pointing distinctly 
in that direction. 

I have some additional remarks to make upon the subject of glaciation, but 
these I shall defer to a subsequent paragraph. 

3. Till or Boulder-clay. — The till or boulder-clay of the Faeroe Islands closely 
resembles the similar deposit which occurs in the hilly and mountainous districts 
of Scotland. We found it in a great many places, generally as little local 
patches, sheltering in the lee of roches moutonnees and projecting rocks ; at 
other times spreading more continuously over low ground, and covering the 
beds of gently-sloping and wide valleys. Not infrequently it occurs along 
the margins of fiords, where the hills retire, and the coast-land is low. It 
varies much in thickness, but seldom exceeds 15 feet, and generally it is much 
thinner. In the neighbourhood of Thorshavn, it is a hard, tough, dark brown 
deposit, stuck full of blunted stones and boulders, some of which were well- 
striated. This was the case especially with some of the bigger stones. The 
same deposit of till showed here and there an irregular layer of earthy gravel of 
the usual character. The clay ranged in thickness from a few feet up to six 
yards ; and here and there contained blocks of basalt that measured 10 and 12 
feet across. Along the shores of the bay it rests upon a glaciated surface, and 
the same is the case with the till at Hoyviig, which is of a dark brownish blue 
colour. I noticed till also in Kolfaredal and at Westmannshavn and Saxen. 
Thin sprinklings were observed at various places between Eide and the foot of 
Slattaretind ; and at Andafiord the low cliffs along the shore are formed of a 
very hard, dark greyish blue till with angular and blunted stones — some of the 
larger ones showing striae. This till rests on a striated surface of dolerite. 

Another good exposure of till occurs on the shores of Boroevig fiord, close 
to Klaksvig. It contains intercalated lenticular beds of fine tough brown stone- 
less, laminated clay and sand, as shown in Plate XV. fig. 19. The till is of the 
usual character, but very few of the included stones show faint striae ; they are 
of the common blunted subangular shape. At Giothe there is a good deal of 
till, and irregular sheets of it appear here and there along the course of Skaale- 
fiord, as at Siov, Strendre, and Glibre. Again in Sandoe considerable depths 
of till fill the bottom of the valley that opens into the sea at Sand. The deposit 
in this valley is more than 20 feet thick, and is well exposed along the course 
of the stream. We noticed till in Suderoe in many places, but more particularly 
in Trangjisvaag valley, and at Ordevig. In the former it occupies the whole 
bottom of the valley, and it also shows in various places along the northern 

vol. xxx. part i. 2 P 



250 DR JAMES GEIKIE ON 

shores of the fiord. Intercalated lenticular beds of gravel, clay, and sand 
occur here and there. 

The distribution of the till and the mode of its occurrence harmonised com- 
pletely with the appearance presented by the glaciated rocks ; it lay either upon 
low undulating grounds, or was closely packed together behind rocks, whose 
abraded and ice-worn faces were quite destitute of any such covering. All the 
stones and boulders in the till were of local origin, and in the many exposures 
which we examined we never saw a single fragment which might not have been 
derived from the islands themselves — all consisting of basalt-rocks and tuff, and 
chiefly of the former. This till, we had no doubt whatever, represents the 
ground-moraine, or moraine defond of the old ice-sheet that covered the islands. 
On steep slopes and in situations which must have been exposed to the full force 
of the ice-flow not a scrap of till appeared. 

4. Erratics and Morainic Debris. — Large angular blocks of basalt-rock are 
of common occurrence throughout the islands, but in some districts they are 
more conspicuous than in others. In the vicinity of Thorshavn they are specially 
numerous, and many of them attain a large size, measuring occasionally upwards 
of 20 feet across. They occupy positions which preclude the possibility of their 
having fallen or rolled down from the hills, and as they are now and again 
associated with coarse morainic debris, I do not doubt that they have been 
deposited in their present positions during the melting of the ice-sheet. An 
excellent exposure of morainic debris, consisting of earth and angular fragments 
of all shapes and sizes up to large blocks, may be seen on the road that leads 
out of Thorshavn on the way to Kirkeboe. Large blocks are frequently seen 
dotting the hill- slopes in greater or lesser numbers throughout all the islands, 
and not a few may have had a glacial origin, but in many cases such isolated 
blocks and morainic debris can hardly be distinguished from the loose fragments 
which are disengaged even now by the action of frost, and rolled down from the 
upper parts of the hills. It is noticeable, however, that while perched blocks 
are quite absent from hill-tops which give no evidence of glaciation, they are 
yet often scattered abundantly over the surface of high ground which has been 
glacially abraded. This is well seen upon the ridge between Giothe and Skaale- 
fiord, where isolated erratics are sprinkled about upon the moutonne'e surface. 
But even in the valleys I found true morainic delbris less plentiful than might 
have been expected. In Kolfaredal, for example, only a slight sprinkling 
occurred, and in many cases this debris might quite well have resulted from 
the shattering by frost of the rocks in situ. This was particularly well seen 
in Saxen valley, where immediately below the lake the bottom of the valley is 
filled with what appear at a first glance to be hummocks of morainic matter. 
I found, however, that many of those hummocks were mere knobs and bosses 
of basalt-rock screened and masked by their own ruins. Whether some of the 



THE GEOLOGY OF THE F^ROE ISLANDS. 251 

debris and loose blocks that were scattered about might not have had a true 
morainic origin, it was impossible to say, but from the presence of some blunted 
and evidently glaciated stones, this seemed highly probable. True terminal 
moraines, however, were noticed in the valley that leads down to Westmanns- 
havn, on the way over from the head of Kolfaredal. At Andafiord in Osteroe, 
I found the till overlaid by true morainic ddbris and shingle, and all the low 
ground was sprinkled with large erratics which could not have rolled to their 
present positions, but are of unquestionably glacial origin (see Plate XV. fig. 20). 
At Klaksvig also may be seen erratics and morainic gravel and sand overlying 
till (see fig. 19). In Suderoe, as in the northern islands, numerous loose erratics 
appear, but morainic ddbris seems to occur in mere superficial sprinklings, and 
never, so far as I saw. formed definite mounds. So far as my observations 
went, well-marked terminal moraines appeared to be quite exceptional, but 
much of the loose angular debris and earth which are scattered over the bottoms 
of the valleys and the lower grounds are doubtless of morainic origin. It is 
needless to add that not a single erratic or loose fragment of rock foreign to the 
islands was observed. 

5. Lake-Basins.— All the lakes, with one or two exceptions, occupy true 
rock-basins. None, as I have already mentioned, attains a large size. They 
are somewhat numerous, and were formerly more abundant, for not a few 
appear to have been silted up, and are now represented by little sheets of 
alluvium and cakes of peat. As all the lake-basins we visited present the same 
kind of features, a few only need be specially referred to. 

The col between Kolfaredal and the valley that takes down to Leinum is 
only 259 feet (79 metres) above the sea. In the Leinum valley, a short distance 
below the water-parting, occurs a little lake (Mjavatn), and about half a mile or 
so further down is a second and larger lake (Storevatn). Mjavatn is divided 
by a cone de dejection of detritus brought down into the valley by a torrent 
escaping from the hills on the north side. The surface of the lake is only 12 or 
13 feet lower than the water-parting — the height over the sea being 75 metres. 
The lake is about quarter of a mile or so in length, and it appears to be shallow. 
Yet it is a true rock-basin, since the stream at its lower end flows over rock. 

Storevatn is three-quarters of a mile or more in length, and half a mile or 
so in breadth, and seems to be deeper than Mjavatn. At its lower end it is 
hemmed in by rock, over which the stream flows. The height of this lake we 
found to be 63 metres, or 207 feet. Roches moutonnees occur at the foot of both 
lakes, but only a thin sprinkling of angular debris is scattered over the valley. 
There are no true moraine mounds. Storevatn lies at the base of tolerably 
steep hills — the one on its north-west side being glaciated all over. The lake 
originally extended up the valley to the north-west, and must at one time have 
been nearly two miles in length. 



252 DR JAMES GEIKIE ON 

Storevatn evidently owes its origin to the grinding action of the ice that 
flowed from the heights between Qvalvig and Leinum. This ice was deflected 
by the mass of the hill called Saaten, and would necessarily erode a hollow 
where the opposition to its flow was greatest. I think it is also highly probable 
that in late glacial times this hollow, which had been excavated at a period 
when the ice was thickest, would be occupied by a local glacier. Mjavatn 
appears to be likewise due primarily to the excavating action of the ice-sheet. 
It lies quite close to the low col or water-parting at the head of Kolfaredal, 
which, when the ice first began to stream down the slopes of the island, was 
probably a more marked feature than it is now. The water-parting may then 
have formed a rocky barrier against which the ice that flowed across into 
Kolfaredal would press with great force. Here then another rock-basin would 
be formed, which, however, would tend to become obliterated as the rock-barrier 
continued to be lowered by the grinding of the strong ice-current that passed 
across it. In late glacial times the hollow may have been to some extent 
modified by the action of a small local glacier. 

Another interesting rock-basin is that in the valley of Saxen. The surface of 
this lake is 22 metres or 74 feet above the sea. The valley in which it lies is 
wide, and comparatively flat-bottomed — the position of the lake being shown in 
Plate XV. fig. 21, which is a diagrammatic longitudinal section of the valley. 
The present stream has cut deeply into the old valley-bottom, and now flows 
some 140 feet or so below its general level. The lake-basin appears to have 
been excavated in the bed of the old valley which rises some 60 feet or so above 
the surface of the water. I saw no trace of old water-levels, which would 
indicate a former greater height for the lake. But its borders are strewed with 
so much debris fallen from the heights above, that any such traces might well 
be obliterated. I think it most probable, however, that the lake-surface was 
never much higher than it is, but that the deepening of the outlet went on con- 
temporaneously with the excavation of the rock-basin, and consequently that the 
deep gully in which the stream now flows is not entirely of post-glacial origin. 
It is worthy of note that the water-parting of the valley in which the Saxen 
lake-basin occurs is a mere flat col like that of Kolfaredal. The glacier to which 
it owes its origin did not therefore head in lofty, steeply-inclined valleys and 
corries, but was rather a thick flattened mass of ice that gathered deeply over 
the low-lying col, and seems to have flowed both to north-west and south-east. 

Besides these larger lakes, which occupy nearly the whole breadth of the 
valleys in which they occur, very many lakelets lie in cirques. The great cirque 
of Howe, for example, is studded with lakelets. I counted upwards of twenty, 
varying in breadth from 50 or 60 feet up to several hundred yards. They are 
all true rock -basins — the bed of the great cirque in which they lie being also 
abundantly covered with well-marked roches moutonnees. In the numerous 



THE GEOLOGY OF THE F^SROE ISLANDS. 253 

smaller cirques that occur in Suderoe and the northern islands, lakes are of 
almost invariable occurrence. Now and again, however, such lakes have 
become silted up, and are replaced either by flats of alluvial detritus and peat- 
moss, or by quantities of rough debris which have tumbled from the surround- 
ing precipices. All these cirque-lakes are unquestionably of glacial origin — the 
cirques themselves having been originally formed by the action of springs and 
frost, and subsequently deepened and excavated by small local glaciers and 
glaciers remanies, as I shall describe more fully in the sequel. 

The rock-basins visible in the islands themselves have thus been excavated 
under varying conditions. Some, as we have seen, have been hollowed out 
when the ice was at its thickest. They belong to the period of general glacia- 
tion, while many are due to local glacial action, which has likewise in some 
cases modified the results produced by the erosion of the ice-sheet. In the 
great amphitheatric cirque of Howe, we see how innumerable small rock- 
hollows may be scooped out by the action of one and the same ice-flow. 
Doubtless, most of these little basins owe their origin to some accidental 
circumstance. In some cases, perhaps, the rock has yielded unequally, owing 
to more abundant jointing or to differences in mineralogical composition and 
petrological structure. In other cases the basins may simply indicate localities 
where the ice, owing to the form of the ground, was enabled to exercise greater 
intensity of erosion. 

VI. Origin of the Valleys and Fiords : Subaerial and Glacial Erosion. 

1. Forms of Valleys. — I have already made brief reference to the various 
forms assumed by the valleys, and must now describe these a little more fully. 
For this purpose I shall select one or two examples which may be taken as 
typical of all the others. 

The great cirque-valley of Howe in Suderoe is one of the finest to be met 
with in any of the islands. It opens upon Howe Bay on the east coast of 
Suderoe with a breadth of nearly one mile. Its bottom is gently undulating, 
being studded with clustering roches moutonnees. Following the stream 
upwards we are suddenly brought to a cliff or wall of rock at a distance of 
rather more than a mile from the sea. This wall circles round the valley, and 
appears to form its head. But when it is surmounted we find a second broad 
cirque-like valley bounded in like manner by steep cliffs, above which other 
cliffs rapidly succeed, rising tier above tier to the summits of the bare rugged 
mountains. This upper cirque-valley is wider than the lower one, and its 
bottom is covered with roches moutonnees, in the hollows amongst which occur 
the numerous lakelets of which I have already spoken. What chiefly im- 
presses one is the great width of the valley relative to its length. From the 
sea to the head of the upper cirque is just some three miles ; yet the width of 



254 DR JAMES GEIKIE ON 

the valley at its upper termination, at the base of Borgaknappan, is nearly one 
mile and a half (see Plate XIII. fig. 1). 

The geologist, crossing from Saxen to Tiornevig in Stromoe, will traverse 
two very characteristic valleys. The stream which comes from the north 
falls into Saxenfiord down a succession of steep cliffs. When we ascend these 
cliffs, which are perhaps 400 or 500 feet high, we find ourselves in a flat- 
bottomed valley bounded by encircling cliffs of dolerite, above and beyond which 
another similar but shorter cirque-valley appears. Between this upper plateau 
and the valley of Tiornevig the water-parting is reached at a height of 1693 
feet (516 metres). From this point the ground descends rapidly towards the 
north-east into another cirque-shaped valley, the sides of which consist of a 
succession of narrow plateaux, so that the stream descends by leaps from one 
stage to another till it reaches the sea at Tiornevig (see Plate XIII. fig. 2). 

Similar features characterise nearly all the valleys. They descend in a 
series of platforms, which vary in number, breadth, and relative height, but 
invariably present the same features. Those I have now described agree in 
this respect that they all have well-defined water-partings ; to get from one 
valley over into the other we must ascend and descend several hundred feet. 
But there are other valleys, the heads of which coalesce, as it were, so that we 
pass from the one into the other over a low flattened col or water-parting. 
Such valleys form the great hollows which I described in a previous section 
as crossing some of the islands from sea to sea. The lofty rock-barriers which 
must at one time have separated the heads of these valleys have been demo- 
lished. The measurements we obtained in Kolfareclal will illustrate the 
general character of those remarkable hollows. From Kollefiord we found 
that the valley rose to the water-parting (259 feet) with a mean inclination of 
13° or 14°. The water-parting itself is low and flat, and it was hard to dis- 
tinguish any culminating point. The descent on the other side of the water- 
parting is very gentle, the fall being only some 50 feet or so for the first two 
miles. After this the sea is reached at Leinum in less than a mile — the fall 
being of course more than 200 feet in this short distance. 

2. Fiords. — The soundings upon the chart prove that the long fiord which 
separates Stromoe from Osteroe, occupies the bed of two submerged valleys 
with a low separating col, over which there is shallow water. This col occurs 
in the narrow part of the sound between Nordskaale and Ore, and the sound- 
ings show that from this point the water deepens both towards north-west and 
south-east. The fiord is shallower at its mouth near Eide, where there are 5^ 
and 9 fathoms of water, than it is at and above Haldervig, where we get depths 
of 18 to 30 fathoms. The southern part of the fiord has not been sounded, but 
it is probably the deeper of the two sections. Many of the other sounds 
between the islands are apparently of the same nature as that just described — 



THE GEOLOGY OF THE FJEROE ISLANDS. 255 

such as Qvanna Sund, Svinoefiord, Harald Sund, Kalsbefiord, Leervigsfiord, 
Westmannsfiord, &c. An elevation of 200 or 300 feet would probably suffice 
to run the sea out of all these fiords and sounds and convert them into valleys, 
communicating with each other across low level cols. They would, in short, 
resemble the long hollows that have already been described as crossing the 
islands from shore to shore. Conversely, were the islands to be submerged for 
200 or 300 feet, new sounds and fiords would make their appearance, and 
Stromoe would be cut up into three separate islands, Osteroe into two, Boroe 
into two, and Suderoe into three, while the present fiords would then cover all 
the low grounds at their origin, stretching back into those broad amphitheatric 
cirques which are so prominent a feature in the configuration of the Fseroe 
Islands. 

Although the fiords are never very deep, they yet, as we have seen, resemble 
those of Scotland and Norway in this respect, that they are shallower at or 
near their mouths than somewhat further up. Unfortunately, the soundings 
indicated upon the chart are not numerous, and reliable details are thus want- 
ing. But the fishermen assured us that it was certainly true that the fiords 
were deeper above than below their entrances. The soundings between 
Osteroe and Stromoe show that Sundene at least has this character ; and it is 
interesting to observe that the deepest portion of that fiord occurs just where 
it should if the depression owed its origin to the grinding action of the ice 
that flowed towards the north. There appears also to be a deep excavation in 
the sea-bottom to the north of Naalsoe, comparable to the deep hollows that 
are met with along the inner margin of the Outer Hebrides and many other 
islands off the Scottish coast — hollows which I have elsewhere shown must be 
attributed to the erosive action of glacier-ice."" 

3. Trend of Valleys and Fiords: Main Water-parting. — One sees at a glance 
that the Fseroe Islands are only the more mountainous parts of a region which 
has been submerged within comparatively recent geological times. And it is 
not difficult to account for the north-west and south-east trend of so many of 
the valleys and fiords. If we draw a somewhat undulating line from north-east 
to south-west between Svinoe and Waagoe, we traverse in so doing the main 
water-shed of the islands, and what must likewise have been the chief water- 
shed when the land stood several hundred feet higher. Now, if we suppose 
the original surface of the land to have been gently undulating, there can be 
little difficulty in accounting for the trend of the valleys and fiords. What is 
now the main water-parting will represent the low undulating water-shed of 
the old table-land — the ground to north or north-west of the line having a gentle 
fall in that direction, while to the south of the line the inclination would be to 

* The Great Ice Age, p. 289 ; Quart. Journ. Geol. Soc, vol. xxxiv. p. 861 ; Trans. Geo! Soe. Glasg., 
vol. vi. p. 161. 



256 DR JAMES GEIKIE ON 

south and south-east. In addition to that main line of drainage, there would 
of course be subsidiary water-partings, such as that which runs down through 
Stromoe between Saaten and the heights that lie to the south of Thorshavn. 
From this water-shed streams would flow east in the direction of Kollefiord and 
Kalbaksfiord, and west into what is now Hestoefiord. 

4. Origin of Main Water-jmrting , &,c. — To what the original superficial un- 
dulations of the old table-land were due it is not so easy to say. Evidence is not 
wanting, however, which seems to show that in some cases the direction of a 
valley may have been determined by the dip of the strata. Thus in Fundings- 
fiord the dolerites dip approximately in the same direction as the fiord or 
E.N.E., and similar appearances were noted in Andafiord, where the beds 
incline to north-east. Again, on the south side of the main water-parting 
nearly all the fiords and valleys run towards the south-east, or approximately in 
the same direction as the dip. On the other hand, it will be noted that some 
of the valleys and fiords trend at right angles to the dijD, particularly in Osteroe 
and Suderoe, while others run in a direction exactly opposite to the inclination 
of the strata. It is possible, therefore, that the coincidence of the dip with the 
trend of certain valleys and fiords may be to some extent at least accidental, 
and that the configuration of the original surface may have been determined 
only in part by the inclination of the bedding. It seems not unlikely, indeed, 
that mere irregularities in the deposition and accumulation of the youngest or 
latest of the trappean strata may have had much to do in producing the primeval 
water-sheds of the old table-land ; so that, while the original streams would in 
many cases flow with the dip of the rock, they might in other cases frequently 
be compelled to run in some different direction. It is even quite possible that 
the strata may have been slightly tilted since the streams first began to carve 
out the hollows which are now land-valleys and sea-lochs. For example, the 
dolerites in the north of Osteroe and Stromoe may have been approximately 
horizontal or even had a slight dip towards the north-west when running water 
commenced to carve out the now submerged valley between Nordskaale and 
Eide. If the subsequent tilting were slowly effected, the erosion of the valley 
might have kept pace with the elevatory movement ; the direction of the drain- 
age need not have been reversed. But we see now such a very small portion 
of the ancient volcanic plateau that it is almost useless to speculate upon the 
various causes, for there may have been many not now apparent, which deter- 
mined the principal water-sheds of the now fragmentary table-land. Mackenzie 
was of opinion that the narrow channels that separate the islands might have 
originated in the destruction of large basalt-veins, removed by denudation in 
the same manner as those smaller veins and dykes which are seen giving rise 
to caves and hollows along the shore-line. The fiords, however, are as we have 
.seen simply submerged valleys, and there is nothing to show that the valleys 



THE GEOLOGY OF THE FMROE ISLANDS. 257 

of the land have been hollowed out along the course of large dykes. Many of 
the small veins which are now exposed to the action of the weather certainly 
crumble rapidly away, and so give rise to more or less deep gullies. We have 
no reason to believe, however, that any dykes actually reached to the original 
surface of the old plateau. And if they did break through that surface they 
must have overflowed, and cooled under circumstances which would necessarily 
produce a rock differing considerably in character from that of the dykes which 
we now see, but probably approaching to that of the dolerites which they 
intersect. 

5. Atmospheric Erosion. — From the appearances presented by the land- 
valleys there can be no doubt that these hollows owe their origin to the action 
of the usual atmospheric agents, aided by the subsequent erosive powers of 
glacier-ice. To make this clear, a few notes on the nature of the subaerial 
denudation now going on seem desirable, and these I shall supplement with 
some remarks on the effects that have been produced by former glacial 
action. 

Although the basalt-rocks of the Fseroe Islands, when freshly exposed, are 
hard and tough under the hammer, yet their composition and structure render 
them peculiarly liable to more or less rapid denudation. Not only are they 
frequently decomposed by the chemical action of the atmospheric forces, but 
their abundant jointing enables frost to act most effectively upon them, while 
the work of demolition is still further aided by the horizontality of their bedding. 
1 have already mentioned the fact that some of the basalts weather more readily 
than others, and that even in one and the same bed there are often great 
differences in this respect. Thus certain basalts show rough irregular surfaces 
— the wacke"-like portions crumbling to earth and sand — while the harder parts 
weather less rapidly, and thus amorphous hollows, and ruts, and shapeless humps, 
knobs, and ridges come to diversify the faces of the rocks in many localities. 
These appearances, however, are best seen in the sea-cliffs. In the inland 
valleys the hollows in the rock are often masked by the fall of debris from the 
ledges above, and are only conspicuous upon very steep cliffs where the loose 
material finds no angle of repose. As some basalts weather more rapidly than 
others, their demolition often leads to the destruction of the harder masses 
that overlie them. The latter are undercut, and by and by large segments 
split off and fall away. But the undermining of the rocks is carried on in the 
most marked manner by springs which frequently issue in abundance all along 
the outcrops of certain beds, particularly when more or less impermeable tuffs 
alternate with the basalts. This action results in a more or less steep cliff,, 
broken by little sloping ledges which are often thickly carpeted with bright 
green mosses, in striking contrast to the bare walls of brown rock above and 
below them (see Plate XV. fig. 23). Not only are the basalt-rocks undermined 

VOL. XXX. PART I. 2 Q 



258 DR JAMES GEIKTE ON 

by the mechanical action of the springs, but they are of course abundantly 
subject to the action of frost, and immense quantities of debris are thus 
showered down the precipices. One can see that the denudation is going on 
rapidly in many places, and thus gradually destroying the glaciated aspect of 
the ground. 

The low dip of the bedding greatly aids the frost and springs in their work 
of destruction, and now and again considerable landslips take place in 
consequence. At Tiornevig in Stromoe we saw the cultivated grounds to a 
large extent buried under masses of debris and large blocks which had fallen 
suddenly only a year or so before our visit, and the evidence of similar landslips 
having occurred in many of the higher valleys of the islands was abundant and 
conspicuous. So rapid, indeed, is the destruction of the mountains that one is 
apt to wonder that the valleys are not more plentifully covered with debris ; for 
the streams are insignificant, and hardly able to carry seaward any large propor- 
tion of the debris showered down the slopes by springs and frosts. Nevertheless, 
some of the streams during floods would seem to overflow wide areas, and 
probably carry seaward no inconsiderable amount of material. They are fast 
silting up the lakes, and many of these have already been obliterated. And 
the same silting-up process is going on in the fiords. Thus Saxen Fiord, which 
was once a good harbour, and could be entered by sloops and other vessels, 
will now hardly admit a small boat. It was not quite low tide when we entered, 
and yet our boat grounded several times, and was only brought to shore by dint 
of vigorous pushing. I have little doubt that could the quantity of material 
carried down by the streams be fairly estimated, we should find it considerably 
in excess of what might have been supposed. 

It is evident that springs and frosts have been among the chief agents in 
widening the valleys. When the streams first cut down into the basalt-rocks, 
they doubtless flowed in more or less deep trenches— the walls of which, under- 
mined by springs and riven by frost, would gradually recede. Moreover, owing 
to the regularity of the bedding, and the low dip of the strata, the retreat of 
the opposing cliff's would be tolerably uniform, so that each valley would tend 
to retain a somewhat equable breadth throughout. But as the streams in their 
course traversed a series of beds — some of which would yield to denudation 
much more readily than others — the waters would descend in a succession of 
runs and leaps. Each hard bed of trap, which happened to be underlaid by a 
more or less thick band of soft tuff or decomposing earthy amygdaloid, would 
give rise to a waterfall, the crest of which would gradually retreat up the valley 
as the trap continued to be undermined by the rapid wasting-away of its pave- 
ment. It would rarely or never happen, however, that all the waterfalls in a 
valley would retreat at the same rate, and thus one would in the course of time 
overtake another, with which it would coalesce, as it were, to form a higher fall. 



THE GEOLOGY OF THE F^EROE ISLANDS. 259 

This heightened cascade would in like manner gradually retreat up the valley, 
and perhaps would merge with others before it finally reached the steeper 
slopes in which the stream originated. The retrocession of the rock-walls 
however, would not be entirely due to the action of the stream — for the under- 
mining process would be carried on at the same time across the whole breadth 
of the valley by the action of springs and frost. In short, the rock-face would 
recede up the valley in the very same manner as the loftier lateral cliffs, 
between which the water flowed. The Fseroe Islands afford numerous examples 
of every stage in this kind of valley-formation. In Kolfaredal we find the 
valley-bottom rising from the sea-level with a gentle acclivity to the watershed 
— the rock-walls and cascades have disappeared. In other valleys, again, we 
have the bed rising with the same low inclination until it is hemmed in, as it 
were, by a rock- wall over which the water pours from an upper platform. This 
latter in like manner slopes gently upwards until it is suddenly cut off by a 
similar rock-wall, above which a third flat -bottomed course succeeds, terminated 
by another steep face of rock, and so on. The steeper the gradient of a valley 
the more numerous do the transverse cliffs and waterfalls become, which of 
course is only what might be expected from the comparative horizontality of 
the strata. 

Valleys, excavated in the manner described, are necessarily more or less 
cirque-shaped at the head, and similar but usually smaller lateral cirques open 
upon them at various levels throughout their course. These lateral cirques 
have been formed by the locally rapid recession of the valley-cliffs. Here and 
there some particular bed, perhaps pretty high up in the cliff, is more quickly 
undermined than the strata below it, and the upper section of the cliff therefore 
tends to recede more rapidly than the under. In this way small lateral cirques 
are formed which collect the tribute of the springs and send down their cascades 
to the main stream. 

6. Former Greater Rainfall. — Now the recession of all those rock-walls and 
cliffs results in the production of enormous quantities of debris, the accumulation 
of which, if it be permitted, must in the course of time bank-up the precipices and 
retard their waste. And this is precisely what is taking place. The streams are 
unable to carry away all the loose material which is brought down the slopes by 
the action of springs and frost. It is evident, therefore, that the time must arrive 
when these slopes will become more or less masked or concealed under their own 
ruins. Even now one may see the process far advanced in many mountain- valleys 
— great curtains of debris hanging from the higher parts of the hills, and sweep- 
ing in long trains down to the low grounds, where they gather upon the bottoms 
of the valleys, and are often left undisturbed by the streams, confined as so many 
of these are to narrow post-glacial trenches. It is clear, then, that there must 
have been a time when the rainfall in the Fseroe Islands was more considerable 



260 DR JAMES GEIKIE ON 

than now — a time when the streams flowed in sufficient body to flood their 
valleys, and to prevent the undisturbed accumulation of debris-banks at the 
base of their cliffs. It is quite impossible that the valleys could have been 
excavated to their present breadth by the small streams of to-day — even with 
all the aid of springs and frost. These streams are now busied in digging 
narrow trenches in the flat bottoms (as shown in Plate XV. fig. 22), forming as 
it were valleys within valleys. 

7. Glacial Erosion of Valleys, &,c. — There are many appearances, however, 
which cannot be explained by aqueous erosion even on the supposition that the 
rainfall was formerly more excessive. The width of many of the cirques, the 
form of the valley-bottoms, the presence of rock-basins, and other phenomena 
all testify to powerful glacial erosion. The rounded and somewhat undulating 
contour of the valley-bottoms, and the smoothed and bevelled appearance of 
the cliffs are conspicuously glacial. The valleys have been glacially deepened 
and widened, and the harsher features which the cliffs must have presented in 
preglacial times have thus been softened down. The demolition of the rocky 
cols between two valleys is also unmistakably the work of the ice. As with 
the valleys so with the amphitheatric cirques, large and small alike, and whether 
forming the head of a valley or opening abruptly upon a valley from a mountain- 
slope — all have been considerably modified by glacial action — many containing 
rock-margined lakelets. When local glaciers occupied these cirques, the 
recession of the cliffs by which they are surrounded would proceed at a rapid 
rate — the debris, instead of gathering at their base and forming a protecting 
mantle, being carried outwards and drifted over the rock-ledges to some lower 
glacier, or glacier remanie, upon the surface of which they might be carried 
down to sea. 

8. Weathering of Glaciated Surface. — But throughout all the islands the 
features impressed by former intense glacial erosion are now, as I have said, 
being more or less rapidly effaced. Roches moutonnees are breaking up and 
disappearing ; ice-worn cliffs are being chipped and shattered by frost ; great 
taluses of debris are accumulating at the foot of scaur and precipice ; rock- 
basins are being tapped and silted up ; streams are digging deep gullies in the 
flat glaciated bottoms of cirques and valleys ; and thus erelong the characteristic 
ice-worn outlines will vanish, and those features which must have characterised 
the islands in early preglacial times will come more and more prominently into 
view. 

0. Limited accumulation of Till upon Land. — When the islands were 
enveloped in their ice-sheet, the action of frost would be confined to such ridges 
and hill-tops as projected above the mer de glace, while severe glacial abrasion 
would go on below. This abrasion, carried on doubtless during a prolonged 
period, resulted in the more or less complete removal of all great banks of debris 



THE GEOLOGY OF THE F^EROE ISLANDS. 261 

that cloaked the valley-slopes, — in the bevelling and rounding-off of basalt- cliffs 
and ledges, — in the deepening and widening of cirques and valleys, the levelling 
of valley-bottoms, the reduction of low-lying water-partings or cols, and the 
excavation of rock-basins. It is in accordance with all that we know of the 
glacial phenomena of Scotland, Norway, and Switzerland, that the material 
produced by glacial abrasion should not have collected in any great quantity 
under the ice. The gradients, as a rule, are too steep, and comparatively little 
till, therefore, was accumulated in the valleys, the great mass having doubtless 
been rolled out to sea, and spread over the sea-bottom — part of it probably 
being carried away by the icebergs that broke off from the terminal front of the 
ice-sheet. 

10. Direction of Ice-flow and Extent of Ice-sheet. — The undulating lines which 
I have described as indicating the primeval water-shed of the old table-land 
also mark out the centres from which the mer de glace flowed. The long sound 
that separates Osteroe from Stromoe brimmed with ice which flowed in two 
directions. North of Nordskaale the movement was northerly, while south of 
the shallow part of that sound the ice held on a southerly course. So thick 
was the mer de glace that its upper strata flowed across Kollefiord and 
Kalbaksfiord, and even overwhelmed Naalsoe. All Stromoe south of Kalbaks- 
fiord appears to have been smothered in ice flowing to south-east, and I believe 
that the direction of the flow in Hestoefiord was the same. Sandoe was also 
overwhelmed, nor can there be any doubt that the ice which covered it was 
continuous with that which cloaked all the islands to the north. Of these last 
it is enough to say that so far as our observations went, they appear to have 
been glaciated invariably in the direction of the principal fiords. A glance at 
the map, indeed, will show that the ice streamed outwards everywhere from the 
dominant high grounds. 

The' appearances in Suderoe are extremely interesting, inasmuch as they 
prove that the ice of that island, although for the most part strictly local, and 
flowing east and west from the chief heights, was yet connected with the mer de 
glace of the northern islands. This is shown by the fact that glacier-ice has 
passed up Qvalboefiord across the island to the west coast. That ice could not 
have come from Suderoe itself, but from a thick glacier-mass occupying the bed 
of the sea between Suderoe and Sandoe. In other words, the ice-sheet of the 
northern islands must have coalesced with that which covered Suderoe. This is 
not astonishing when we remember that the mer de glace of Stromoe must have 
reached a thickness, at what is now the sea-level, of 1500 or 1600 feet. So 
thick a mass could not have floated off in the shallow water (20 to 40 fathoms) 
that separates Suderoe from the northern islands. But the ice that streamed 
south from Stromoe and Osteroe was thicker even than these depths imply. 
To the north of Naalsoe we get a depth of 120 fathoms which must have been 



262 DR JAMES GEIKIE ON 

filled up of course before the ice could overflow that island. This indicates a 
mass of ice not less than 2200 or 2300 feet in thickness. In Suderoe, again, the 
upper surface of the ice attained a height of 1400 feet, which we may take as 
the thickness of the stream that flowed out of Trangjisvaagfiord and Howebugt. 
We cannot wonder then that the shallow seas which separate the Fseroe Islands 
were completely filled up, nor that the outflow of ice from Suderoe should have 
been controlled by that coming towards it from the north. The direction of 
the stride and roches moutonnees at the head of Qvalboefiord agrees with that of 
the glaciation in the west of Sandoe, and seems to point to a general movement 
of the mer de glace towards the south-west. Probably all the ice that lay to 
the west of Skuoe and the two Dimons flowed in this direction, while that which 
lay to the east of these islands participated in the south-easterly movement. 

How far out to sea the ice-sheet extended can only be conjectured, but 
judging from the thickness it attained upon the islands, and in the sounds and 
fiords, it may well have reached what is now the 100-fathoms line, where it 
would break away in bergs. Like the greater triers de glace of Europe and 
North America, it tells a tale of excessive evaporation and precipitation, and 
one ceases to marvel at the thickness attained by the ice-sheets of those vast 
continental regions when one sees the indisputable evidence of a very consider- 
able sheet of glacier-ice having covered even so limited an area as that of the 
Faeroe Islands. 

11. Origin of Erratics and Morainic Debris. — The large erratics which 
are scattered over hill-tops and hill-sides were doubtless deposited by the mer 
de glace during its final dissolution ; and the erratics and loose morainic debris 
that occur in the valleys mark out, in like manner, the gradual disappearance 
of those local glaciers that still occupied the hollows of the islands after the 
higher grounds had been relieved of their icy coverings. I have mentioned 
the fact that valley-moraines are much less numerous than one might have 
expected, when the shattery character of the igneous rocks is kept in mind. 
It is true that considerable quantities of loose morainic-like debris are fre- 
quently scattered over the bottoms of the valleys, but well-marked mounds of 
morainic origin appear to be of rare occurrence. We saw some in the valley 
leading down to Westmannshavn, and some of the mounds in the Saxen valley 
may be moraines, but such of them as were exposed in section proved to be 
roches moutonnees buried under their own ruins. It is difficult to account satis- 
factorily for this scarcity of moraine mounds, more especially when we remember 
that in other countries, such as Scotland, North of England, Wales, Ireland, 
Norway, Switzerland, &c, where the rocks as a rule are less easily acted upon 
by the weather, distinct valley-moraines are yet more or less abundantly met 
with. The following considerations, however, may help to explain the apparent 
anomaly. 



THE GEOLOGY OE THE F^EROE ISLANDS. 263 

1st. Long after the general mer de glace had become reduced to a series of 
small isolated glaciers, it is probable that snow and neVe' may have continued 
to cover the land so as to protect the rocks from the excessive waste which 
they now experience. The superficial moraines, therefore, need not have been 
very extensive. 

2d. Again, we must remember that the conformation of the ground, unlike 
that of the Alps, would not favour the preservation of large valley-glaciers 
after the snow-fall had become less, and the snow-line had retreated to a 
higher level. The ice which formerly occupied all the valleys of the Fseroes 
could only have been sustained by the copious precipitation of snow over the 
whole land-surface. Bat when the line of perennial snow had risen to 1000 
feet or thereabout, only a few local glaciers would probably exist at the heads 
of the valleys, while glaciers remanies would be distributed along the flanks of 
the valleys at all those points where torrents and cascades presently appear. 
There would be no large trunk glaciers formed by the union of considerable 
lateral glaciers, as in the valleys of the Alps. Thus there would be a general 
absence of terminal moraines in the middle of a wide valley, for the superficial 
debris would be distributed chiefly along the flanks of the glaciers and in front 
of the small glaciers remanies. Moreover, the moraines thus formed would tend 
to become obscured by the gradual accumulation upon them in post-glacial 
and recent times of rock-rubbish detached by the weather from the cliffs above. 

3d. But the principal reason for the apparent scarcity of valley-moraines 
was probably the continuous and comparatively rapid dissolution of the ice 
after the snow-line had retreated several hundred feet above the sea level. 
The ice would appear to have made no long pauses in the valleys, as the 
ancient valley- glaciers of Britain did, but melted gradually and continuously 
away. The distribution of the morainic material in sheets over the beds of the 
valleys seems to point to the destructive action of water escaping from the 
dissolving glaciers. The detritus thus formed has much the character of that 
loose aggregation of coarse shingle, earthy grit, and boulders, which is known 
in Switzerland as " Alpine diluvium." 

VII. Marine Erosion. 

The erosive action of the sea is admirably exhibited along the shores of 
the islands, and more especially upon those parts of the coast-line that face 
the open ocean. Everywhere the cliffs are undermined and eaten back, the 
rate of erosion varying according to the character of the rocks at the sea-level. 
The rapid recession of the cliffs is aided not only by the soft and decomposing 
character of so many of the basalt-rocks and tuffs, but also by the abundant 
jointing of the rocks ; and springs and frost are evidently as actively 
engaged along the sea-cliffs as they are upon the steep slopes that overlook 



2(34 DR JAMES GEIKIE ON 

the valleys. But the influence of jointing upon denudation is certainly most 
marked in the sea-cliffs. Nowhere can this be seen to better advantage than 
along the magnificent shore-line of Stromoe between Westmannshavntiord and 
Myling. The cliffs there are nearly vertical, and show broad, bare, plane 
faces, that look often as if they had been only freshly fractured or sliced. 
Towards the top they are more rugged, and grass grows on all the little ledges, 
giving the appearance to the cliffs of having been sprinkled with green tufts. 
The upper parts of the cliffs are often riven by the frost into peaks, spikes, 
and spires. This great rock-wall, I may add, ranges in height from 1200 feet 
or so up to 2277 feet. Between Muulen and Saxen occur some splendid stacks 
called the " Drangar " (isolated or lonely ones). Some of these are not less 
than 400 feet in height. They taper upwards to sharp pinnacles, and one of 
them, called " Toskuradrangar," which forms a long wall running parallel to 
the cliff, is pierced by a lofty portal. Long vertical master-joints are con- 
spicuous in the cliffs at irregular intervals, and give rise to hollows and caves. 
All the caves we saw between Muulen and Saxen were worked out either 
along the lines of such joints or in vertical dykes of basalt. They occur in 
all stages. Here one sees the process just beginning, — a little cleft only a foot 
or two in width, and a few feet in height. There again one observes another 
which shows a somewhat greater breadth and height, the height almost in- 
variably exceeding the breadth. Sometimes as many as twenty caves can be 
counted in the space of a quarter of a mile or less, varying in extent from a 
few feet in height and breadth to large caverns 20 to 50 feet in height, and 
somewhat less in breadth, which penetrate the cliff for some considerable 
distance. The master-joints just referred to seem to cut the cliffs at right 
angles to their trend, and they are crossed by another set of great joints which 
have the same direction as the coast-line. Thus when the sea has undermined 
the cliff to a certain extent, the strata eventually give way and great segments 
are sliced off along the lines of jointing that run parallel to the shore. These 
joints are, of course, not so conspicuously visible as those which cut them at 
right angles. Nevertheless, they were seen again and again on the sides of 
projecting headlands, and the clean fracture presented by the faces of the cliffs 
themselves clearly indicate their presence. The large sea-stacks or "drangar" 
seemed to me to owe their origin to the destruction of caves which had been 
hollowed out along both lines of jointing, the long wall-like sea-stack called 
" Toskuradrangar " being evidently defined by joint-planes. This well-marked 
cross jointing has also given rise to the remarkable indentation in the cliffs 
which occurs a little south of the " drangar," where the cliffs retire so as to 
form a kind of marine amphitheatre about 60 or 70 yards in diameter, and 
surrounded by nearly vertical precipices rising to some 1200 or 1400 feet. 
Similar joints are well seen along the coast of most of the other islands. 



THE GEOLOGY OF THE FxEROE ISLANDS. 265 

They are finely displayed in those of Waagoe, Osteroe, Sandoe, Skuoe, and 
Suderoe, and I have no doubt that they have played a most important part in 
determining the trend of the coast-line where that faces the open ocean. In 
the quieter fiords the action of the sea, although frequently well-marked, is of 
course not so conspicuous. 

Basalt-dykes almost invariably give rise to caves. This is due not so much 
to the decomposition of the basalt (which is generally a harder and less easily 
decomposed rock than the bulk of the bedded strata) as to the minutely 
fissured or jointed character of the intruded rock. The dykes are traversed 
by many long joints parallel to their direction, and by innumerable cross joints, 
so that they fall an easy prey to frost and the battering of the waves (see 
Plate XIV. section fig. 14) One of the most beautiful dyke-caves is that to 
which I have already referred, the " Hole under Kjetle" (Plate XV. fig. 16). 

From the fact that caves occur solely at the sea-level and are nowhere seen 
inland, we gather not only that they are of purely marine origin, but also that 
no part of the Fseroe Islands has been submerged within any late geological 
period. We searched everywhere for evidence of the former presence of the 
sea, but failed to find the slightest proof that the islands were ever smaller 
than they are now. This is in keeping with what my brother and Messrs 
Horne and Peach have observed in Orkney and Shetland, and with what I have 
noticed in the Outer Hebrides. The belief amongst the inhabitants, indeed, 
is that the land is sinking, but the facts mentioned by them in support of this 
view do not appear to be satisfactory. Thus the shallowing of Saxen Fiord 
might be entirely due to the action of the Saxen stream, aided by the tide 
itself. There are several considerations, however, which lead me to believe 
that the islanders are probably right in their conjecture. The present coast- 
line does not appear to me to be very old. Had it been of long standing I 
should have looked for more evidence of excessive marine erosion. We know 
that along the present coast-line of Scotland a terrace of marine erosion is 
frequently visible. Our land has stood so long at its present level, that the 
waves have cut back the cliffs for some distance, so that at low tide a platform 
or terrace of rock is more or less exposed at their base. Now the rocks of the 
Freroe Islands are being denuded much more rapidly than those parts of the 
Scottish coast-line to which I refer ; and had the land remained at its present 
level for any prolonged period, we might certainly have expected to meet with 
such rock-platforms in the Faeroe Islands more or less abundantly. But the 
cliffs seem generally to shoot down at once into deep water, and only in a 
few places were sea-stacks in any number. A submerged rock-platform 
diversified with numerous stacks, some of which peer above the sea-level, 
occurs along the west coast of Suderoe, but this is apparently the only island 
which is thus reef-fringed. And even these reefs are perhaps too deeply 

VOL. XXX. PART I. 2 R 



2G() DR JAMES GEIKIE ON 

submerged to have been cut down by the waves at their present level. I 
should infer that, if the islands are not now sinking, they have within recent 
times experienced some degree of depression. 

VIII. Peat and Buried Trees. 

The low grounds and gentler slopes in the islands are often coated with 
thick turf or peat, which is extensively dug for fuel. It varies in thickness 
from two to six or eight feet ; and here and there, in what appear to be the 
bottoms of old lakes, it may possibly be thicker. So far as my observations 
went, the Sphagnacew appeared to form a smaller proportion of the peat than 
is the case in Scotland. At the bottom, roots and twigs of brushwood are 
frequently present, and in some places they are quite abundant. The largest 
pieces we saw were not more than an inch thick, but we were informed by 
the people, who were digging the turf in Osteroe, Stromoe, Sandoe, Suderoe, 
and other islands that sticks as thick as one's wrist were common ; and at 
Eide in Osteroe, the merchant there told me he had seen them taken from the 
peat near Eide Vatn as thick as his leg. I could not determine the species 
of wood with any certainty, but the pieces I saw were probably juniper and 
birch. No brushwood now exists anywhere in the islands, except in the 
private gardens and enclosures at Thorshavn, yet the evidence supplied by the 
peat makes it certain that the lower parts of the Faeroes must formerly have 
been pretty well clothed with brushwood and small trees. This was of course 
in post-glacial times, when the islands were probably of considerably greater 
extent, and enjoyed a climate more suited than the present to the growth of 
arboreal vegetation.""" 

* I have elsewhere endeavoured to show that the Fseroe Islands were connected with Scotland in 
post-glacial times. See Prehistoric Europe, p. 518. 




THE GEOLOGY OF THE FJEROE ISLANDS. 267 



IX. EXPLANATION OF PLATES. 

Plate XIII. 



Fig. 1. Eepresents a horizontal section, drawn along the line A.B., upon the map, across the 
Island of Suderoe. The horizontal and vertical scales are the same. The horizon of 
the thin coal got at Dalbofos is shown by a thick line ; but it must be understood 
that the coal itself has been found only at one place. The beautiful columnar basalt 
of Frodboe is shown at b. 
Fig. 2. This is a diagrammatic representation of the form of the ground between Saxen and 

Tiornevig in Stromoe. 
Fig. 3. Is a similar diagrammatic section across Osteroe, between the sea at Eide and the head 
of Fundingsflord. These sections illustrate a very common feature in the valleys of 
the Fteroe Islands, many of which descend in successive platforms. 
Fig. 4. Is a horizontal section, drawn to same scale as fig. 1, across the north end of Suderoe 

along the line CD. 
Fig. 5. Sketch-section of an exposure of strata at the coal- workings of the Praestefield (Suderoe), 
at a height of 147 metres above the sea. The coal occurs in two layers, and is sepa- 
rated by 4 to 6 inches of shaly clay. The shales overlying the coals are interlaminated 
with lines and thin seams of glance coal, which are sometimes as much as 3 or 4 
inches thick. There is a good deal of iron pyrites in the shales and coal-lines above 
the coal. The reference letters are as follows : — 
d. Tumbled blocks and superficial debris. 
t. Till or boulder-clay. 

i. Nodules of coarse grey arenaceous clay-ironstone ; 1 to 3 inches in diameter. 
s. Dark shales and clays with lines and layers of glance coal and common coal ; 

about 8 feet thick. 
c\ Glance coal ; 8 inches. 
d. Dark shaly clay or clunch ; 2 to 8 inches. 
e 2 . Coal ; 2 feet. 

g. Space covered with grass and debris ; tufaceous shales. 
a. Dark anamesite. 

The dip is north-east at 5°. 
Fig. 6. Sketch-section of strata at Praestefield (Suderoe), near the place shown in fig. 5. The 
spot is 161 metres above sea-level; the dip of the strata is towards north-east at 3°. 
The reference letters are as follows : — 
t. Till or boulder-clay. 

s. Shales and clays, about 12 feet thick, with layers and seams of coal, as in 
section fig. 5. The coal-lines are lenticular, and are most abundant in the 
shales immediately overlying the thick coals ; towards the top of the section 
they are less plentiful. 
i. Ironstone nodules in shales. 
c 1 . Coal; 8 inches. 
d. Shale or clay ; 4 inches. 
c 2 . Coal ; 1 foot 2 inches. 



268 DR JAMES GEIKIE ON 

Fig. 7. Sketch-section of strata at Dalbofos (Suderoe) ; 607 feet above sea-level : — 

a. Anamesite ; earthy and decomposing into rude spheroids below. 
s. Fine grained greenish tufaceous shales, with fine mudstones. 

s i. Nodules of coarse ironstone in greenish tufaceous shales. 

b. Tufaceous, agglomeratic, scoriform upper surface of anamesite. 

c Probable position of thin coal seam (3 inches) described by Forchhammer ; 
not now seen in place. 
Fig. 8. Section of upper part of sea-cliff at Syd i Hauge, on west side of the Prsestefield 
(Suderoe). Height of coal about 440 feet above the sea. From mouth of mine to top 
of overlying anamesite 50 or 60 feet. 

ft 1 . Dull, dark blue, fine-grained anamesite ; showing amygdaloidal band towards 
middle of the bed ; decomposing spheroidally below. 
s. Shattery dull green tufaceous shales with streaks and lines of coal. 

c. Coal, about 3 feet ; seamed with layers and streaks of clay or shale. 
ft 2 , ft 3 . Anamesites of same character as a 1 . 

t. Line of red palagonitic tuff. 

Plate XIV. 

Fig. 9. Diagrammatic sketch-section of part of sea-cliff at Frodboenypen (Suderoe). 

ft 1 . Bed of anamesite, 50 or 60 feet thick, showing lines of amygdaloidal cavities. 
s. Brown shales and clay ; 15 or 20 feet thick. 
e. Coal, about 1 foot 6 inches ; with parting of shale or clay. 
s i. Shattery brown mudstones resting on brown ferruginous shales, with some 

nodules of ironstone ; 20 to 30 feet thick. 
ft 2 . Anamesite, decomposing spheroidally. 
Fig. 10. Sketch-section of basalt-veins ; northern shore of Qvalboefiord, near Qvalboe. 
d d. Basalt-veins. 
t. Bed tuff. 
Fig. 11. Sketch-section, sea-cliff, Stromoe, near the Drangar. 
ft. Bed tuff. 
b b b. Amygdaloidal bands or zones in dolerite. 
c c. Non-amygdaloidal dolerite. 
ft". Overlying bed of dolerite. 
Fig. 12. Sketch-section ; sea-cliff near Eidevig, Fundingsfiord. 
d 1 ft 7 * ft 73 . Beds of dolerite. 

ft. Zone of wacke-like amygdaloidal rock, forming part of bed d 2 . 
t t. Beds of red tuff. 
Fig. 13., Sketch-section ; sea-coast near Qvitenses, Stromoe, showing thinning out of red tufftf, 

between beds of dolerite d. 
Fig. 14. Sketch-section ; sea-cliffs west coast of Stromoe, a little south of entrance to SaxeD 
Fiord, bb Dykes and veins of basalt traversing beds of dolerite, with separating 
layers of red tuff. Dykes are much jointed both longitudinally and transversely. 
Their thickness varies from a mere thread up to 12 or 14 feet, //are joints travers- 
ing the cliffs and giving origin to caves c 1 . c 2 shows a cave excavated in the line of 
a dyke. 

Plate XV. 

Fig. 15. Ground-plan of a dyke with veins, near Saxen, Stromoe. 

Fig. 16. ^Diagrammatic ground-plan of cave (" Hole under Kjetle ") at southern entrance to 

Saxen Fiord, c a shows the south opening into the cave, and c b the north entrance. 

The arcade is excavated along the lines of cross dykes which are represented in the 

engraving by the cross hatching. 



THE GEOLOGY OF THE F^EROE ISLANDS. 269 

Fig. 16. h. shows the appearance of the dyke as seen in the roof of the cave. 

Fig. 17. Sketch-section of dyke; sea-cliff, west coast of Skuoe. The cliff shown in the sketch 

is about 600 feet high. 
Fig. 18. Hill-slope near Westmannshavn, to show glaciated contour — the sharp edges of the 

basalt-beds being smoothed off. 
Fig. 19. Sketch-section, sea-shore, near Klaksvig, Boroe. 
p. Peat. 

m. Boulders, shingle, gravel, and sand. 

t. Tough hard dark blue till, with subangular glaciated stones. 
c. Tough brown laminated clay and sand, with no stones. 
Fig. 20. Sketch-section ; sea-coast, Andafiord, Osteroe. 

a. Coarse unstratified shingle, and rough earthy grit with large boulders. 

Morainic debris. 

b. Very hard, dark greyish blue till, with subangular, angular, and blunted 

stones and boulders — some of the larger ones showing striae. The deposit 
rests on a striated pavement of blue dolerite ; the striae point out to sea 
or E. 40° S. 
Fig. 21. Diagrammatic longitudinal section of Saxen Valley. I lake; v bottom of old glacial 

valley ; s s level of present stream (Giogveeraa) ; F level of fiord. 
Fig. 22. Diagrammatic section across Saxen Valley ; v bottom of glacial valley ; s bed of 

present stream ; d dtibris from subaerial waste of cliffs. 
Fig. 23. Diagrammatic section to show denudation of basalt-rocks, d, caused by springs creeping 
between these and intercalated layers of tuff, t. Talus of debris forming at base of 
cliff shown at r. 

Plate XVI. 

Geological Map of Faeroe Islands. The general features of the geology of the Faeroe Islands 
are correctly indicated upon the map which accompanies Dr Forchrammer's paper (" Om 
Faeroemes geognostiske Beskaffenhed," Kongl. danske Vidensk. Selsk. Skrifter, 1824), of which 
the present is almost a reproduction so far as the delineation of the coal-crops is concerned. 
The dips of the strata, the dykes and veins, and the direction of striae and roches moutonnees 
are from my own observations. The coal-crops also differ somewhat from the lines given by 
Forchhammer, and from those on the more recent map of Suderoe by Professor Johnstrup. 



VOL. XXX. PART I. 2 S 



}zz7z*s. 2&y. <5oc.2?c&jl'. 



ft 






V 



> 






r 



& 






nz,3zzj a 6v&^zzr. 












% 






V 








%■- 


\__ 






ir 




< \ 


V 

5S 


^"-v < 


/ 






Hf- 


/ 


X 


f*i 


% 




s 








^ 



■V ^ "^'rS 




J Baniolamow, E&rf 



rf«s Jfay. J'ocJEduc'. 



~v r oi^zz:ttz&y_zrp: 




J". B ardiojjin •■ \r. E iru r 



hz7uJRoy. <5bcJ?dz'nZ 



ToZ ZWr MateXfr 






:';F:;:.:-' , -'- : >^:- :: -<;-;; ; i::: 













<^ 



I'W' 



• p 



0-,0\ 






\ If V » o 



% 

* 



J. BartlioloinTw; Edinf 




I 







^ J H $ ^ £ I 



I 




( 271 ) 



X. — Researches in Contact Electricity: Thesis for the Degree of Doctor of 
Science. By Cargill G. Knott, D.Sc. Communicated by Professor 
Tait. (Plate XVII.) 

(Received July 23d ; revised October 27th, 1879). 

At the surface of separation of any two different substances in contact, 
there exists in general an electromotive force tending to maintain a certain 
difference of potential between them. This principle, established for metals by 
Volta in 1796, has been extended by later investigators to other substances, 
including liquids and gases. From these early researches of Volta/" and the 
later more elaborate inquiries of Kohlrausch, t Hankel, \ and Gerland, § 
there have been deduced certain fundamental laws, which have been fully 
corroborated by the recent work of Clifton, || and Ayrton and Perry. IF If, 
of a number of conductors set serially in contact, the difference of potential 
between each successive pair is quantitatively estimated and reckoned positive 
or negative, according as the first member of the pair is at a higher or lower 
potential than its successor, then the difference of potential between the first 
and last members of the chain is equal to the algebraic sum of the potential 
differences between the successive contiguous pairs. Should the series be made 
up of simple conductors, the potential difference between the extremities is 
quite independent of the nature, number, and order of the intervening com- 
ponents, and is, indeed, equal to the difference obtained by direct contact of 
these extreme members. Hence, in a circuit composed of such substances 
(metals for example) and kept at a uniform temperature throughout, the sum 
of the differences of potential existing at the various surfaces of contact taken 
in order all round the circuit is zero. The resultant electromotive force is 
therefore also nil, and no current can exist. This result of experiment is in full 
accordance with the recognised principle of the conservation of energy, there 
being in these circumstances no source from which the current could derive its 
energy. Should the contact-chain, however, consist partly of compound or 
chemically decomposable conductors, the potential difference between the 

* Annales de Chimie, vol. xl. p. 225 (1801); also Wiedemann's " Galvanismus," vol. i. §§ 1-7 
and 14. 

t Poggendorff's Annalen, vol. lxxxii. p. 1 (1851), and vol. lxxxviii. p. 465 (1853). 
% Ibid., vol. cxv. p. 57 (1862), and vol. cxxvi. p. 286 (1865). 
§ Ibid,, vol. cxxxiii. p. 513 (1868). 

|| Proceedings of the Boyal Society (London), vol. xxvi. (1877). 
H Ibid., vols, xxvii. (1878), and xxviii. (1879). 
VOL. XXX. PART I. 2 T 



272 DR CARGILL G. KNOTT ON 

extremities may, and frequently does, become a function of the intermediate 
structure, and is then no longer equal to the direct contact-difference between 
the extreme members.* With a circuit including such materials in its com- 
position, the resultant total electromotive force does not in general vanish, so 
that the existence of a current is possible and necessary; and the energy of this 
current is derived from the energy of chemical combination, which is the one 
aspect of the accompanying action, whose other aspect is the decomposition 
of the compound conductor. Except such chemical action were possible no 
current could be generated ; so that, probably, the possibility of chemical 
action, and the non-vanishing character of the resultant electromotive force in 
the circuit, are necessarily co-existent phenomena. Such, at present, seems to 
be the most complete theory of the voltaic cell. 

Although no current exists in a circuit of simple conductors maintained at 
a uniform temperature because of the mutual balancing of the contact forces, 
it is possible to cause a current to flow by heating or cooling one of the 
junctions, and thereby destroying the equilibrium of the contact forces. The 
energy of the thermo-electric current so obtained is a partial transformation of 
the energy which was originally expended in unequalizing the temperature of 
the system. Apparently, then, the contact-difference of potential between two 
metals or other simple conductors depends upon the temperature — a conclusion 
verified in a very remarkable way by consideration of the Peltier effect, or 
reversible thermal effect, produced by the passage of a current across the 
j unction of two different metals. By an application of the dynamical theory of 
heat, Sir William Thomson t proved that this evolution or absorption of heat at 
the junction, according as the current flowed in one or other direction, indicated 
an electromotive contact force, acting against the current when heat was 
evolved, with the current when heat was absorbed. In other words, because of 
the difference of potential at the junction, the current has to do work when 
passing in one direction, and has work done upon it when passing in the other — 
giving rise respectively to an evolution and absorption of heat. From considera- 
tion of the principle of dissipation of energy, Professor TaitJ has developed a 
formula for this electromotive contact force, expressing it as a parabolic 
function of the temperature ; and this theory has been indirectly verified by a 
long series of experiments upon the thermo-electric properties of metals. 

With a view of testing by direct contact experiments the variation with 
temperature of the contact-difference of potential between dissimilar metals, I 
undertook the experiments whose results form the subject of this thesis. It 
must be premised, however, that any positive result cannot be regarded as due 

the papers of KoiiutAuscn, Hankel, Clifton, and Ayrton and Perry, cited above. 
t Transactions of the Royal Society of Edinburgh (1851). 
J Ibid. (1870-71). 



RESEARCHES IN" CONTACT ELECTRICITY. 



273 



only to the metals ; for, as pointed out by Professor Clerk Maxwell,* Volta's 
electromotive force of contact is in general greater than that indicated by the 
Peltier effect, and sometimes of opposite sign — a discrepancy to be accounted 
for by the fact that in direct contact experiments there is always a film of con- 
densed air or other gas between the metals when they are in so-called contact, 
and that possibly the chief effect " must be sought for, not at the junction of 
the two metals, but at one or both of the surfaces which separate the metals 
from the air or other medium which forms the third element of the circuit." 
After a few preliminary experiments I concluded that direct contact of the 
surfaces under investigation was a sufficiently accurate and constant method 
for indicating any appreciable change which might occur. It was found neces- 
sary, however, to keep the surfaces continually polished in a particular manner, 
since they gradually altered their surface condition when exposed to the action 
of the air — a fact formerly established by Hankel. t Previous to any discus- 
sion of the results obtained, it is advisable first to give a description of the 
apparatus and method of experiment. 

Of the two metallic surfaces which were the subject of experiment the 
lower was the upper surface of a flat cylindrical flask-shaped vessel, which 
rested on an insulated stand in electric connection with one pair of opposite 
quadrants of a Thomson Quadrant Electrometer. The temperature of this 
surface was determined by the temperature of the water contained in the flask. 
Three such flasks were used — one of iron, one of zinc, and the third with the 
one flat surface copper and the other tin. The diameters of the plane faces, 
the thicknesses of the flasks, and their volume capacities, are as follows : — 



Flask. 


Diameter in 
Millimetres. 


Thickness in 
Millimetres. 


Volume in Cubic 
Millimetres. 


Iron, 
Zinc, 
Copper and Tin, 


128 
131 
129 


17 
16 
16 


198,000 
168,000 
196,000 



The upper plate of the condenser was a disk of approximately the same 
area as the lower, on which it pressed during contact by its own weight. It 
could be readily adjusted by screws to a practically accurate parallelism with 
the lower plate, and had only one degree of freedom — an up-and-down motion 
directed by a pin and guiding slot. It depended from the brass top of a 
cylindrical glass case which surrounded the insulated stand and flask on all 
sides, if we except the small aperture through which the internal arrangement 
was put in connection with the electrometer. Great care was necessary in dry 

* Electricity and Magnetism, vol. i. § 249. 
t Pogg. Ann., vol. cxxvi. p. 286 (1865). 



274 DR CARGILL G. KNOTT ON 

weather to avoid rubbing, and thereby electrifying this glass case, which during 
tlu' experiments had to be repeatedly removed, so that the temperature inside 
might be observed and the surfaces repolished. The upper and movable plate 
of the condenser was connected with the other pair of electrometer quadrants, 
which were put to earth and kept constantly at zero potential. In all cases 
the plates were brought into direct surface contact, and the deflection on the 
electrometer scale caused by the charge left on the insulated flask and the 
connected quadrants, when the upper plate was withdrawn to a height of five 
inches, was taken as the quantitative estimate of the difference of potential due 
to the contact of the surfaces. These opposed surfaces were polished with 
emery paper, and dusted with dry chamois skin. The polishing was effected 
manually, the surface to be polished being held for the time in one hand, and 
the emery paper in the other, and the two rubbed vigorously together for a 
quarter of a minute or so. After being thus polished the surfaces were dusted 
and reset in as short a time as possible, an interval of about fifteen seconds 
elapsing between the polishing of the second surface and the first contact of 
the two plates. 

In the first series of experiments the condenser-plates remained almost 
always in contact, being separated only when a reading was to be taken, or 
when the surfaces had to be repolished and the temperature of the water in 
the flask observed. The upper disk was then virtually at the same temperature 
as the lower. Headings were taken in groups of five at a time, the interval 
between each reading being conditioned by the swing of the electrometer 
mirror, which, under the action of the bifilar suspension, had of course to come 
to rest, or nearly so, before its indications could be of any value. After each 
group of readings the surrounding glass case was removed, the temperature of 
the cooling water observed, the surfaces repolished, and the whole arrangement 
re-adjusted precisely as before. On the whole, the five consecutive readings 
of any group were very consistent considering the difficulties besetting elec- 
trometer measurements of contact forces, and were sufficiently so in all but 
a few very exceptional cases to warrant the belief that, during the two or three 
minutes necessary to make the complete set of readings, comparatively little 
change took place on the surfaces. From theoretical considerations I was led 
to try iron and copper as likely metals to give positive results. In this I was 
not disappointed ; but the difficulty of drawing any sure conclusion from the 
indications so obtained, or in any way deciding between the claims of the 
various possible explanations which might be given to account for the facts, 
induced me, after four months experimenting, to conduct the inquiry on a 
different, and what turned out to be an improved, principle. In these earlier 
experiments it is to be particularly observed that the two surfaces were at any 
instant both at the same temperature; in the later experiments the tempera- 



RESEARCHES IN CONTACT ELECTRICITY. 



275 



ture of the lower surface only was made to vary, so that the surfaces were 
generally at different temperatures. By the former method it was found that 
the difference of potential between polished iron and polished copper fell off 
by at least T ^ath °f its original value for a rise in temperature of 1° C. Many 
series of experiments were made with these two metals, and each day's results 
gave the same general indication ; although, as might have been expected from 
the nature of the inquiry, it is hardly possible to deduce from them any definite 
quantitative law. 

The genera] results of eleven series of experiments are given in the follow- 
ing table. The first column represents the lowest temperature for which 
readings were taken ; the second gives the electrometer deflection for that 
temperature due to the electrification by contact of the lower surface ; the 
third indicates the like deflection for the higher temperature ; the fourth 
registers that higher temperature ; and the fifth notes the percentage average 
decrease of the deflection for unit increase of temperature. 



1 

Lower 

Temperature. 


Deflection. 


Deflection. 


Higher 
Temperature. 


Percentage 
Decrease. 


7°C. 


70 


50 


30°C. 


1-32 


14 


60 


35 


40 


1-60 


13 


78 


58 


45 


•83 


10 


77 


56 


45 


•78 


18 


110 


80 


50 


•85 


20 


93 


76 


41 


•87 


23 


110 


85 


50 


•84 


16 


110 


60 


48 


1-42 


16 


110 


85 


35 


1-22 


20 


112 


91 


38 


1-04 


16 


112 


90 


36 


•98 



The first four experiments give smaller readings than the last seven — a 
discrepancy easily accounted for by the change of circumstances occasioned by 
removal to another room, and a refitting of the surfaces. Yet, that in such 
altered circumstances the average percentage temperature-variation of the 
deflection should be so consistent throughout, argues strongly in favour of the 
reality of such a variation. A like series of experiments was commenced with 
a zinc surface substituted for the copper or under surface ; but, though there 
were indications of a somewhat similar variation, these were too vague to 
admit of any definite deductions being made. The same was true of the 
aluminium-zinc pair. In this mode of experimenting, however, it was impos- 
sible to determine how much of the resultant variation of a given pair was clue 
to the action of a particular component, or how far this variation depended 
directly upon the change of temperature, or indirectly through consequent 
material alteration of the surfaces — through oxidation, for example. 



276 DR OARGILL G. KNOTT ON 

These considerations led me to abandon my first method of experimenting ; 
and in the modified method finally adopted, the temperature of the upper plate 
of the condenser was kept constant, while the temperature of the lower was 
made to vary. This required the contact to be instantaneous, so that only one 
reading could be taken between each preparation of the surfaces and observa- 
tion of the variable temperature. During this interval the upper plate was 
laid upon an iron slab, and thus kept at the temperature of the room ; and 
just before the apparatus was reset for observation the temperature of the 
lower surface was noted, and both surfaces were polished and dusted as usual. 
The first experiments were made with two iron surfaces, which, after sufficient 
polishing at the ordinary temperature of the air, gave no deflection on separa- 
tion after contact. The lower surface was then heated up to 70° or 80° C. in 
the manner formerly described, and then allowed to cool, while at rapid 
intervals instantaneous contacts were made with the upper surface, each 
contact being made as soon after polishing as possible. In this way I found 
that iron hot was strongly negative to iron cold, and apparently more negative 
the higher its temperature — in other words, the difference of potential between 
iron and iron increases with the difference of temperature, being zero when 
the temperature difference is zero. A glance at the representative curve 
(Diagram, fig. 2) shows the nature of this change. The six different symbols 
represent six different curves, five of which give the results of as many 
independent experiments, while the sixth (represented by the circle and dot) 
is the average curve formed by the combination of the others. Each point on 
any one of the five primary curves, is, as far as possible, the mean of five 
consecutive readings — a method of reduction which recommended itself as 
giving the most probable value for the mean contact. Each point of the 
final mean curve is obtained by taking the average of all such points as lie in 
the same temperature decade. Subjoined are the tabulated results of these 
experiments, the upper row of each of which gives the temperature of the 
lower condenser plate, and the lower the corresponding deflection on the 
electrometer scale. 

Experiment I. {February 27, 1879). 
(Curve symbol .). 

Temperature (in degrees C), . . . 53'8 49-4 45-4 31 23-1 

Deflection, 21-8 21 17"4 10-8 7-2 



Experiment II. {February 28). 

(Curve symbol + ). 

Temperature, 61'9 40'7 31'6 22*4 

Deflection 317 25-2 147 77 



RESEARCHES IN CONTACT ELECTRICITY. 277 





Experiment III. {March 5). 








(Curve symbol x ). 






Temperature, 


64-6 47-3 


31 


22- 


Deflection, 


37*7 26 

Experiment IV. (March 6). 
(Curve symbol V ). 


15-2 


12 


Temperature, 


66-5 36-7 


34-2 


25 


Deflection, . 


47-5 17 


16 


3 



Experiment V. (March 13). 

(Curve symbol A). 

Temperature, . . . 604 55'5 38 30 23-2 

Deflection, ... 37 36 20 9-3 6 

The reduced means for curve VI. (symbol 0) are as follows : — 

Temperature, . . . 631 54-6 45-7 33'4 23-2 
Deflection, . . . 38-5 28-9 22-4 14-9 T2 

The temperature of the room, and therefore of the upper surface, was 
12°C, at which point then the curve should meet the line of temperatures. 
The mean curve is obviously best represented by a straight line, whose tangent 
of inclination to the temperature line is— 76, expressed in diagram units. 

In seeking for an explanation of the results of these experiments, we 
must not neglect the possible effects due to surface oxidation, or to the 
change in density of the gas condensed upon the metallic surface. If the 
negative character of heated iron to cold iron disappeared on the cooling of 
the former, then the effect must be the result of some temporary change 
accompanying the heating — such for example as the mere change of temper- 
ature, or the driving off of the condensed gases at the higher temperature, 
or of both causes combined. Experiment, however, clearly proved that 
the originally heated surface, when cooled to the temperature of the colder 
surface, retained its strong negative characteristics with no appreciable diminu- 
tion ; from which it would appear that the observed phenomena are to be 
attributed mainly to a permanent change of surface condition depending upon 
the temperature to which that surface has for a brief period been subjected — 
probably to oxidation. It was also found by trial that no appreciable increase 
in the deflection corresponding to a given temperature resulted when a 
considerable interval of time was suffered to elapse between the polishing 
of the heated surface and the making of contact between it and the upper 
and colder surface. Whether the instantaneous contact was made fifteen 
minutes (the usual interval) forty minutes, or sixty minutes after polishing, 



278 DR CARGILL G. KNOTT ON 

the electrometer deflection was, as far as the method admitted of judg- 
ing, the same. Probably after a longer lapse of time than that here speci- 
fied, a change might become manifest — such a change as Hankel long ago 
established for iron and other metals at the ordinary temperature of the 
air. In order to compare this time-variation of surface condition with the 
temperature-variation established above, I made a series of observations, 
at sufficiently distant intervals of time, of the deflections produced by con- 
tact and separation of two iron surfaces, one of which was kept constant 
by polishing, while the other was permitted to vary, by being simply left to 
itself. Both were initially polished to be the same electrically — a state of 
affairs evidenced by the absence of any effect on the electrometer when the 
two plates were separated after contact. Readings were first taken at intervals 
of five minutes, then at intervals of ten minutes, fifteen minutes, and finally 
at half hour intervals. Each number in the following table is the mean of 
five readings taken in rapid succession within the lapse of one minute. 



Experiment X 


(May 20). 






(Kg- 1, 


V). 




- 


Time (in minutes). 






Deflection 


(iron against iron) 














5 








-11 


10 








-14 


15 








-15 


20 








-16 


30 








-18 


45 








-19-4 


75 








-20 



The curve corresponding to these numbers is given in the diagram (fig. 1, b). 
In its main characteristics it is very similar to an ordinary curve of cooling, 
and is markedly dissimilar to the curve which represents the temperature- 
variation of surface condition. Curves a and c on the same diagram indicate 
the corresponding variations for copper and aluminium respectively. The 
copper was electrified by contact with iron, both surfaces being allowed to 
vary ; and the real time -variation of the copper was obtained by properly 
introducing the known time-variation of the iron. The aluminium was elec- 
trified by contact with polished zinc, to which it was originally positive, but in 
the course of half an hour became as strongly negative. The contacts were 
instantaneous, and except immediately before the taking of a reading the 
surfaces were kept far apart. The tabulated values for these metals are given 
below, the chemical symbol for each metal being employed to represent the 
corresponding surface, and the suffix^ signifying that the surface to which it 
i- suffixed was kept polished and therefore constant. 



RESEARCHES IN CONTACT ELECTRICITY. 279 

Experiment XI. (May 21). 
(Fig. 1, a). 

Time (in minutes). Cu | Fe Cu | Fe p ' Cu | Cu p 

( = Cu | Fe+Fe | Fe„). 

-63 -63 

2 -61 -67 - 4 

7 -58 -70 - 7 

17 -57 -73-7 -10-7 

47 -55 -75-5 -12-5 



Experiment XII. (May 9). 







(Fig- 1, 


c). 




Time (in minutes). 






Al | Zn„ 


Al | A 









+ 18 





10 






- 7-4 


-25-4 


20 






-14-6 


-32-6 


30 






-16-2 


-34-2 


45 






-24-2 


-42-2 


60 






-24 


-42 


90 






-36 


-54 


1350 (observe 


id next 


morning) 


-47 


-65 



In experiment XL, the second column contains the observed values ; the 
third is calculated from it by adding to each number the corresponding number 
from the iron curve ; and the numbers of the fourth column are obtained from 
those of the third by subtracting from each the first number, which gives the 
deflection due to polished copper and polished iron. In experiment XII. there 
is no column corresponding to the second column of experiment XL, since the 
zinc surface employed for comparison was kept constant throughout the 
experiment. The corresponding curves for zinc and tin are not represented on 
the diagram because of their great proximity to the iron curve. In the course 
of an hour the change on the zinc was only 6 per cent, greater than the 
corresponding change on the iron ; while in forty minutes there was no appre- 
ciable difference in the changes on the tin and iron surfaces. 

The gradual character of the change here indicated is of special value in the 
present inquiry, as I hope to bring out in the final conclusions to which I have 
been led. Meanwhile it is advisable to give the results of the experiments on 
the temperature-variation of the other metals which I investigated. Though 
not so full and satisfactory as the results for iron, these later researches all 
indicate the same general facts — as may be gathered from the following tables 
for copper hot against iron cold, both surfaces being polished with emery paper 
immediately before contact. 

VOL. XXX. PART I. 2 U 



280 



DR CARGILL G. KNOTT ON 



Experiment VI. {March 24). 

(Fig. 3 ; symbol . ). 

Temperature. Deflection (Cu 

62° C. -66-6 P 

57 -68-2 

52 -64-5 

48 -61-5 

44 -57-3 

32 -54 

24 -52 

12 -50 



Fe„) 



Experiment VII. (March 25). 
(Fig. 3 ; symbol x ). 



Temperature. 
70° C. 
55 
43 
30 
23 
12 



Deflection (Cu p 
-69 
-65 
-62 
-52 
-47 
-47 



Fe P ) 



The conditions under which these experiments were made were the very same 
as those under which the temperature-variation of the iron was investigated. 
The representative curve is shown in fig. 3, all the points clustering approxi- 
mately round a straight line whose tangent of inclination to the temperature 
axis is — '39, measured in diagram units. Hence it appears that the tempera- 
ture-variation of copper is smaller than that of iron, and that consequently, 
since the iron is the more positive metal, the difference of potential between 
iron and copper falls off as the temperature of both is raised — a result already 
obtained in the earlier experiments (see page 275). 

Zinc was the next metal which came under investigation. At first it was 
electrified by contact with aluminium, kept polished at a constant temperature. 
This latter metal, however, is not very suitable, on account of its proneness to 
rapid change in time as evidenced by its curve on the diagram (fig. 1, c). 
Nevertheless the same negative growth of the heated metal was indicated, and 
more self-consistent results were obtained by contact of zinc hot with zinc cold, 
both polished as usual. The numbers are as follows : — 



Experiment VIII. (March 28). 
(Fig. 4 ; symbol • ). 

Temperature. Deflection (Zn„ 

63°-8C. -78 

46-5 -66 

34 -64 

21-8 -56-3 

10 -40 



Al„) 




RESEARCHES IN CONTACT ELECTRICITY. 281 



Zn p ) 



Experiment IX. 


{April 4). 






(Kg- 4; 


symbol x ). 




Temperature. 






Deflection (Zn p 


65° C. 








-42 


45 








-22-5 


42-7 








-19-2 


40-6 








-17-7 


38-8 








-18-5 


28-6 








-8 



In the diagram (fig. 4), two lines are drawn, each representing one of the 
above experiments. The dotted line is that which best agrees with the readings 
of experiment VIII., the points on the curve of which are represented on the 
diagram as " dots." The curve-points of experiment IX. are entered as crosses, 
and they all lie very near the continuous line drawn on the diagram. The 
tangent of inclination of this line is — '9, expressed in diagram units. 

Apparently, then, zinc varies more rapidly with temperature than iron ; and 
hence, since zinc is the more positive, the contact difference of potential between 
zinc and iron falls off, as both are simultaneously raised in temperature ; 
a result in accordance with the indications of the earlier experiments with 
zinc and iron when both were made to vary similarly in temperature. This 
suggested the possibility that the more positive metal might be subject to 
the greater temperature-variation. According to this hypothesis, tin, which 
occupies in the electromotive series a position intermediate to zinc and iron, 
should give a correspondingly intermediate line for its temperature-variation. 
It was impossible, however, with the means I had at my disposal, to arrive at 
anything like a quantitative result for tin. Not having at the time another tin 
surface, I was compelled to make use of either zinc or iron as the other con- 
denser plate ; and, as both of these gave large deflections with tin, the readings 
were wild and unsatisfactory. No experiment gave even self-consistent results; 
and no two of them had much in common — except the undoubted characteristic 
which indicated a similar " negative growth " with rise of temperature of the 
tin surface. 

As already noticed, the permanency of this negative-growth with temperature 
increase after the surface is cooled — a characteristic which was established by 
direct experiment in every case — proves conclusively that whatever change in 
the electromotive force of contact of any two of the metals, iron, zinc, copper, 
and tin, may be due directly to change of temperature ; such a possible change 
is quite inappreciable by ordinary contact methods, and is altogether masked 
by changes due to other and secondary causes. In seeking for such causes, we 
must consider the probable alteration with temperature in the density of the 
gaseous film condensed over the metal surface, which alteration, however, is not 



282 DR CARGILL G. KNOTT ON 

permanent on restoration to the original temperature, provided the surface 
has remained the same chemically. Any permanent alteration in the density 
of the condensed gases presupposes, then, a chemical change on the surface ; 
and if there be no such permanent alteration, or if it be insufficient to account 
for the observed phenomena, the last resource still seems to be chemical change, 
to which accordingly we look as the only efficient cause, whether directly or 
indirectly, of the changes observed in the mutual electrical relations of metals. 
This hypothesis is also supported by the known phenomena of time- variation of 
metal surfaces in both their chemical and electrical relations. The electrically 
negative character of unpolished iron, copper, zinc, tin, aluminium, &c, to 
polished iron, copper, zinc, tin, aluminium, &c, is generally attributed to surface 
oxidation ; probably, then, the electrically negative character of polished and 
heated iron, copper, zinc, and tin, to polished but unheated iron, copper, zinc, 
and tin, is to be referred to a similar cause. If so, then the above experiments 
lead to the result that for these metals at least, there is for every temperature 
a definite surface condition which no amount of polishing can alter — a surface 
condition produced most probably by a film of oxide or other similar compound 
over the metallic surface by the action of atmospheric air ; and that, further, the 
surface change due to change of temperature is a direct function of that tem- 
perature-change. This surface state forms within the first few seconds after 
polishing, perhaps instantaneously, and thereafter no appreciable change ensues 
till several minutes have elapsed, when the inevitable time-variation of the 
surface, as depicted in the curves of fig. 1, begins to show itself. Hence it 
would appear that at ordinary temperatures a chemically pure surface of these 
four metals in air is an impossibility; and that the same holds true for other 
metals, even for the so-called non-oxidisable, is a not improbable surmise. In 
this connection it should be remarked that to the eye there was no appreciable 
alteration of surface, no dimming of the bright metallic polish, even after the 
lapse of several minutes. 

The experiments which form the subject of this thesis were made in the 
Physical Laboratory of Edinburgh, during the summer session of 1878, and the 
winter session 1878-79. The apparatus was, for the most part, lent me by 
Professor Tait, whom I here thank for the kindly interest he has evinced in 
my work, and the ever ready advice with which he has aided me. 

Added, May 1881. — As it was just possible in the above experiments that 
the variations of potential observed might be affected by changes in the capacity 
of the condenser, further experiments were made in which any such alteration 
in capacity might be effectively eliminated. The two opposed surfaces of the 
( 'indenser, brought to within a millimetre distance of each other, were put into 
metallic contact by means of external wires. In this way, after the method o 



RESEARCHES IN CONTACT ELECTRICITY. 283 

Kohlrausch, any change in the difference of potential could be measured in 
terms of a Daniell cell. The results obtained fully corroborated the former 
conclusions, as a glance at the following table will show. The first column, 
headed 8 V, gives the variation of potential for a rise of temperature of 1° C. 
expressed in terms of a Daniell cell ; and the second column, headed p, 
indicates the range of probable error in the estimate which was deduced as the 
mean of several distinct experiments. 

8 V p 

Zinc, .... --0028 ±"0003 

Iron, .... --002 ±'0004 

Copper, .... --001 ±-0002 

Tin, .... --001 ±-0002 

It may be noticed that of these zinc gave the most regular results. In 
deducing these numbers it was assumed that the variation varied directly with 
the temperature throughout the range of 60° C. Thus, polished zinc at 20° C. 
gives with polished zinc at 80° C. a difference of potential equal to '168 of a 
Daniell cell — the hotter surface being, of course, the negative surface. 

Many definite results were also obtained for the time-variation for aluminium, 
zinc, iron, and copper. The representative curves were in all cases similar to 
those shown in fig. 1. This being understood, the following numbers indicate 
the difference of potential between the polished metal surface and the same 
surface after twenty-four hours' standing. 

Aluminium, "3 ^| 

' > (in terms of 1 Daniell cell.) 

Iron, -114 

Copper, -086 J 

It was found, however, that different days' experiments gave somewhat varying 
results — the atmospheric conditions as to temperature, humidity, &c, having 
probably some effect. Indications were also obtained in the course of experi- 
ment that this time-variation depended upon the more arbitrary conditions 
under which the varying surface was allowed to vary; whether, for example, 
it was freely exposed to the air, or was left close to the opposed surface; 
whether it was the negative or positive element in the condenser, and such like. 
Where so many possible factors enter, however, it is extremely difficult to 
draw any sure conclusions. 



VOL. XXX. PART I. 2 X 



L 



I:* 

\ 



*i T 



1 1 1 M 1 1 m 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 M ' ri 1 1 1 1 1 1 1 1 ~m — pjy\ — mm [M 1 1 1 1 — n | || 1 1 1 1 1 1 1 || 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ii ' i 


-z=========================^^=^====================================z=====z== =========zz-2b=z:====z=====z==zz: 


/ 7 


, I / 


/ / 


__! f ' / 


J^ / 


-j- A ~-±_ — Ap 4- hi — -L- — -_--2 z — _)__._{_ _X_| — i -j- 


Az zz^ £>t- - + J -JL* — i A 


_i y* 


~f_ r ^ 


[ / 


/ s / 


7 J* ^ t 


t i | /( | 


7 / 


| f -i— - ■-- -j',, / 


x± zz zj xx x : *jl ~ ' ± "" " x ±4: 


M ^ . r x _ . x ^ 5____.' _ 


/! -at / .' i 


_l_ f ^ h x X <z_ .,x __ _ zy. 




i:===="====|=#=="=-|====:=e|E:=:E:|:E|E;^ 


-j ^ ^ 


1 )(-1 . ,..,.._. ....... , ,<.._ _ /_ __ , . ^ 




(. .(. 1 __ _, 1 1_ 


XT X -' - ., - . ' X 


>&■■■'...- ■' - - 


S:±±::^::EE:::-::4::::EE:::::::::::::::::::::::::::::+ = :z:::::l:::::: J 


|_ J , ... ! 


XT ■ T » - TT T — — ■ — j — 


X JZ.Z Xi_ !> ' ^ H 


X X s5 ._ .._ ' >* X 


3 ±_l_ _._ s .' ^ 


zxx xjXxxxx xxxx^ xx: xxxx zz xzxzxxxxx: :: :5jv ± zV 


^ X X^ _ t£tr ± 


->s- - --- - K X X e£ X X "' 






_ _ — __ _ _ 










X X ~ - x X 






I 


__ , X__ ,± _ _ ±IJ .: ^ , .__:__. X X 


* ?■ ;« -^ - -4-1 - -,- -;iy X . JL """ 3;l 2 i. Xt ~+~ 


— 3— =-~A^- x__ + 4Z x it it__ x: ::_:: 


—±.-*z&7.Z7i*-?£r? s X ZC X X X 


*--""' J. X "' 




|ix X ^ H 11 jl 


X - !^- H ^ 1^ <x + X 




a zevpr. . /&/ . 










" -L -t— -1-r- " - 


| r 






I - 






| "I" 




i i ^ z; e zb 




?3 


>». / V 


X " _ : ii 2^ " 


> + / 


1 ,. , , ,|. nj1 - . ,. , , / * 




T.i r " u _ET . ^ EL_ 




y 


Z"Xt _ " i " i 




\ "•? X f" E y X j 


(. .5 Lj * 


X Si 23 ± J / X 






." : ' / 




z:~>~ f J i _ 2 


t x r . ± /,. i/ ... _i i_ 


XX 1 t ^^ + j +■ 


■ - - 1 / 


l r 


~ -H- f- A- *- -t- 




y y 


T j ' 


1 1 1 .j -t- -J- 


X ±. t ^ J4- 


x ~^sl EF «v ^s,!zE j x r yi* EE" 




h : --Z-— J r s- * tz. :i: - Z^ — -::::::-::::!: 


i It-,- k S^ L J - X X 


r t *r j _ ^^ 


t ^ z ± 1 ^x 


x 4- -/ / + 




- -' Xj jl -+- 


i i x xj 2 2 ... x x 


X X _t 1 2 -- - 


— — l_ __ _ 2+ - — -' 


X 1 „ ' y ± X 5- zh -u. 


X ^i j 2 " SI XX X 


r / / > 


r— --j— — 1- 1 f- ^|v Efr r z 


-flX-J-—- -X-Xp^L_J /- ^ S + X_p 


:: = EEES=z|xEE=^:- : " :: " 


' l lr i /^ig^'XT xz 


^^^--- 4h X =p J_____| 



-Zifrr^e. -Ttez-z^w^.^ 






\ 



h 



( 285 ) 



XI.— On Phosphorus- Betaines. By Professor Letts. (Plate XVIII.) 

(Read January 3, 1881.) 

In a paper by Professor Crum Brown and the author on Dimethyl- Thetine 
and its Derivatives* attention was drawn to the analogies which frequently 
exist between elements which have different atomicities, and which are usually 
considered as belonging to different families. The most striking examples of 
such elements are boron and carbon, gold and platinum, and phosphorus and 
sulphur. 

Since the publication of that paper, the author has pursued the subject, and 
his experiments, which have been made with the object of comparing the 
properties of analogous compounds of nitrogen, phosphorus, and sulphur, have 
confirmed the view that the two latter elements are very closely related, and 
that in many cases at least, phosphorus is more nearly allied to sulphur than 
it is to nitrogen. 

In the course of these experiments many facts and considerations relative 
to the three elements have occurred to the author, which he believes have not 
hitherto been presented in a clear and concise form. No doubt, some of them 
have been noticed by other chemists, but he believes that such has not been 
the case with all, and he is therefore induced to give a slight sketch of the 
analogies and differences which the three elements exhibit, before proceeding to 
describe his experiments. 

A Comparison of the Properties of Nitrogen, Phosphorus, and Sulphur. 

If we compare the three elements in the free state, we cannot but be struck 
with the very close analogies existing between phosphorus and sulphur, and 
the great dissimilarity of nitrogen to either. 

Phosphorus and sulphur are solid bodies ; both exist in allotropic modifi- 
cations which are produced by the action of heat on a particular form of each 
element. Nitrogen is gaseous, and so far as is known does not exist in 
more than one condition. 

Again, both sulphur and phosphorus have what is usually termed " abnor- 
mal " vapour densities ; that is to say, in the gaseous state their molecules 
contain more than two atoms. At a sufficiently high temperature, however, 
the molecules of sulphur are dissociated into simpler ones containing two atoms, 

* These Transactions, vol. xxviii. 
VOI,. XXX. PART I. 2 Y 



286 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

and this fact, considering the similarity of the two elements, renders it probable 
that at a sufficiently high temperature the molecules of phosphorus would 
behave in a similar manner.* 

Regarding other physical properties of the three elements, such as specific 
gravity, atomic volume, &c, it is not necessary to say much, as nitrogen, on 
account of its gaseous nature, does not admit of a ready comparison with the 
other two. It may be mentioned, however, that both the atomic weight 
and specific gravity of phosphorus and sulphur are very close to each other, 
and consequently their atomic volumes are nearly identical.f 

Turning now to the chemical properties of the three elements (in the free 
state), we again find a close similarity between phosphorus and sulphur, 
whereas nitrogen possesses scarcely a point of resemblance to either ; for 
whilst the former are characterised by their energetic attraction for other 
elements, nitrogen is strikingly inert, and displays scarcely any tendency to 
enter into combination. 

The great affinity of phosphorus for oxygen needs no comment ; that of 
sulphur for the same element is considerably less, but is still well marked ; 
whilst nitrogen possesses so slight an attraction for oxygen, that its oxides are 
powerful oxidising agents. We have then in phosphorus, sulphur, and nitrogen 
a group of elements which show a regular gradation in affinity for oxygen ; 
and, as we might expect, the affinity of these elements for hydrogen is in 
exactly the reverse order, ammonia being the most stable of their hydrides, 
and phosphuretted hydrogen the least, whilst sulphuretted hydrogen stands 
midway between them. We might perhaps expect from these facts that, as 
ammonia is the most alkaline of all the hydrides, sulphuretted hydrogen would 
be more alkaline than phosphuretted hydrogen ; but this is not the case, for the 
latter has a neutral reaction, and combines directly with hydriodic and hydro- 
bromic acids, whereas sulphuretted hydrogen has a slight, but still a distinct 
acid reaction, and does not, so far as we know, combine with any hydracid. 

The difference observed in the affinity of phosphorus, sulphur, and nitrogen, 
for oxygen and hydrogen, exercises, as we might expect, an important influence 
on the properties of their compounds. Thus most compounds of phosphorus, 
with electro-positive elements or compound radicals, oxidise spontaneously, as 
in the case of phosphuretted hydrogen, many metallic phosphides, and the 

* The author has communicated with Professor Victor Meyer on this subject, who stated that he 
had already made experiments in this direction, and that they indicated a diminution in the vapour 
density of phosphorus at a high temperature. Professor Meyer having thus established his priority to 
any experiments on the vapour density of phosphorus at high temperatures, the author has left the 
matter in his hands. 

t According to Ramsay (Journ. Chem. Soc, 1879), the sp. gr. of sulphur at its boiling-point is 
1-4799, and its atomic volume (in the sense in which Kopp employs the term) 21 '6. The same author, 
in conjunction with Masson (Journ. Chem. Soc, 1880), gives the sp. gr. of phosphorus at its boiling- 
pointe as 1*4850, and its atomic volume as 20-91. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 287 

phosphines ; and even partly oxidised compounds of phosphorus often greedily 
absorb oxygen, and are, as a consequence, powerful reducing agents. 

Similar compounds of sulphur do not as a rule oxidise spontaneously, or if 
they do so the oxidation occurs slowly, as with solutions of sulphuretted hydro- 
gen and metallic sulphides. But oxidising agents easily attack them and 
convert them into oxidised products. Thus sulphuretted hydrogen, by simple 
contact with sulphuric acid, is oxidised to water and sulphur. Organic sul- 
phides (R' 2 S) are converted by treatment with nitric acid into sulphanes (B/ 2 SO), 
and sulphones (R' 2 S0 2 ) ; mercaptans (R'HS) into sulphonic acids (R'HS0 3 ). 

Corresponding compounds of nitrogen show much less tendency to oxidise, 
and only in a very few cases are they capable of directly fixing oxygen ; thus 
in the case of the compound ammonias although oxidised products are known 
(R'NO, R / N0 2 , &c.) they are not produced by direct oxidation. 

These considerations help us to understand the action of reducing agents 
on oxidised compounds of the three elements, and also explain why completely 
different methods must be employed for obtaining their organic compounds. 
A nitro-body is an oxidised compound of nitrogen ; in it the oxygen is only 
weakly held, consequently a reducing agent easily removes it, and usually 
causes the addition of hydrogen.* Consequently an amine is readily obtained 
by the reduction of a nitro-body. Oxidised compounds of sulphur are also 
easily reduced. Thus nascent hydrogen de-oxidises sulphuric, sulphurous, and 
hyposulphurous acids, and converts them into sulphuretted hydrogen, and is 
also capable of converting (certain at least of the) sulphanes and sulphones 
into sulphides. But it is more difficult to reduce an oxidised sulphur com- 
pound than an oxidised nitrogen compound. For instance, nitrate of potash 
is easily reduced to nitrite, and eventually to oxide of potassium by heat alone ; 
whereas sulphate of potassium suffers no change when heated unless a reducing 
agent such as carbon is present ; in which case, however, the oxygen is removed. 

But if we attempt to remove oxygen from an oxidised compound of phos- 
phorus by ordinary reducing agents, we experience as a rule much greater 
difficulty. It is stated that both phosphorous and hypophosphorous acids may 
be reduced by nascent hydrogen,t but phosphoric acid is not affected by that 
reagent, nor is the oxide of a tertiary phosphine. A powerful reducing agent 
acting at a high temperature must generally be employed for the reduction 
of an oxidised compound of phosphorus. 

We can therefore readily understand why phosphines cannot be prepared 
by the reduction of oxidised organic compounds of phosphorus, whilst amines 
are produced by such a process with the greatest ease, and even sulphides are 
formed from sulphines, sulphones, &c, without much difficulty. 

* Not however in all cases, as we see in the preparation of azo-bodies. 
t This statement requires confirmation. 



288 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

Respecting compounds of oxygen and of hydrogen with the three ele- 
ments, it may not be superfluous to point out some of the more important and 
interesting points of resemblance and difference which exist between them. 

As regards their compounds with hydrogen, nitrogen forms a single 
hydride ; sulphur, two ; phosphorus, three. In all three cases the hydride 
containing the maximum of hydrogen is gaseous, and possesses a powerful and 
characteristic odour and energetic properties. All three of these gaseous 
hydrides are decomposed by the spark, and phosphuretted and sulphuretted 
hydrogen are decomposed by heat. Ammonia, however, is more stable. 

As we might expect from the readiness with which both sulphur and phos- 
phorus are oxidised, their compounds with hydrogen are very inflammable, 
whilst ammonia can only be burnt under special conditions. 

The strongest point of analogy between ammonia and phosphuretted 
hydrogen is, that both are alkaline substances, in which respect they are unique 
amongst the hydrides of elements. But the alkaline properties of phosphuretted 
hydrogen are very weak, as it combines under ordinary atmospheric pressure 
with only two acids, viz., hydriodic and hydrobromic acids, and its compounds 
with these are so unstable that they dissociate at ordinary temperatures, and 
cannot exist in solution. 

As before pointed out, phosphuretted hydrogen, in respect of its alkaline 
properties, is intermediate between the strong base ammonia and the faint acid 
sulphuretted hydrogen. In other respects, phosphuretted hydrogen is more 
allied to sulphuretted hydrogen than it is to ammonia. This is especially 
noticeable in its action on solutions of the heavy metals, where it acts either 
as a reducing agent (gold, &c.) or precipitates a metallic phosphide (cadmium 
and copper), or precipitates a mixture of the metal and metallic phosphide 
(mercury). 

Both sulphur and phosphorus form only two well-marked compounds with 
oxygen ; whilst nitrogen, in spite of its slight affinity for that element, forms no 
less than five oxides. 

Phosphorus, as we might expect from its powerful affinity for oxygen, 
combines directly with the maximum quantity of that element ; whilst sulphur, 
when burnt, only forms its lower oxide ; and free nitrogen is not capable of direct 
oxidation, except under special conditions. 

The highest oxides of the three elements resemble each other in being 
volatile solids, and in having a strong affinity for water. Nitric anhydride is 
the least stable, and frequently decomposes spontaneously. Sulphuric anhy- 
dride is decomposed at a high temperature, whilst phosphoric anhydride dis- 
plays a much higher degree of stability. 

If we consider the oxy-acids of the three elements, we see that an undoubted 
analogy exists between sulphuric and phosphoric acids. Both are very powerful 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 289 

acids. Their salts are stable at a high temperature, and in a great many cases 
their solubility is similar. 

Nitric acid cannot be said to resemble either sulphuric or phosphoric acid, 
nor can its salts be compared with sulphates or phosphates. 

There is a distinct analogy between hypophosphorous and hydrosulphurous 
acids, and between phosphorous and sulphurous acids. The first two are 
extremely powerful reducing agents, and to the best of the author's belief they 
are the only substances which precipitate cuprous hydride from a solution of 
a copper salt. Sulphurous and phosphorous acids are also reducing agents, 
but by no means such powerful ones. 

It is rather curious that in this series of acids, so far as their formulae are 
concerned, the only difference between corresponding terms is that all the 
members of the sulphur series contain two atoms of hydrogen, whilst those 
of the phosphorus series contain three. 



H,S0 2 


H 3 P0 2 


H 2 S0 3 


H3PO3 


H 2 S0 4 


w 



There is one point in which sulphur does not resemble either phosphorus or 
nitrogen, viz., in the large number of oxy-acids which it forms. No oxy-acicls 
of phosphorus or nitrogen have been obtained corresponding with hyposul- 
phurous acid or with the polythionic acids. 

Phosphorus and sulphur also agree in their strong affinity for the halogens, 
especially for chlorine, whilst nitrogen has almost no attraction for them. The 
chlorides of sulphur and of phosphorus resemble each other in certain of their 
properties. Thus the higher chlorides of both readily dissociate into chlorine 
and the lower chlorides, and this is especially the case with the chloride of 
sulphur, SC1 4 which dissociates even at ordinary temperature into SC1 2 , or 
S 2 C1 2 and free chlorine. Again, these higher chlorides act upon the hydrates 
of organic radicals, giving their oxychlorides, chloride of the organic radical, and 
hydrochloric acid. 

The two following equations will illustrate this — 

C 6 H 5 -COOH + SCI, = HC1 + SOCl 2 + C 6 H 5 -C0C1 
C 6 H 5 -COOH + PC1 5 = HC1 + POCI3 + C 6 H 5 -COCl 

The lower chloride of sulphur is decomposed by water, with formation of 
hydrochloric and sulphurous acids (and free sulphur) ; and the lower chloride of 
phosphorus is decomposed in a similar manner, with formation of hydrochloric 
and phosphorous acids. 

There is a very striking difference between the three elements in their affinity 
for carbon — a difference that explains several facts which at first sight appear 
anomalous. It is difficult to say whether nitrogen or sulphur has the strongest 



290 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

affinity for carbon; for although, undoubtedly, bisulphide of carbon is obtained 
with greater ease than cyanogen — the compound ammonias (bodies in which 
carbon is directly united to nitrogen) — are so numerous, stable, and so easily 
obtained, that we must accord to nitrogen a very high degree of affinity for 
carbon. Phosphorus, on the other hand, has but a slight attraction for carbon. 
The two elements do not combine directly (so far as we know) to form a com- 
pound analogous to cyanogen, and even the compounds which phosphorus forms 
with hydrocarbon radicals (phosphines) are only obtained with difficulty. 

This striking difference between the three elements explains, in the 
author's opinion (in some measure at least), the curious fact, that whereas both 
nitric and sulphuric acid readily act on a large number of aromatic bodies in 
such a manner that the nitrogen or sulphur becomes directly united to the 
carbon which they contain, phosphoric acid or anhydride is without action on 
them. Considering the analogies which certainly exist, and are always insisted 
upon, between nitrogen and jmosphorus, and also those which exist (but 
are not so commonly insisted upon) between sulphur and phosphorus — we 
should certainly be strongly inclined to predict, if we did not know to the 
contrary, that "phospho" bodies ought to be easily produced by the action of 
phosphoric acid or anhydride on aromatic hydrocarbons. It is almost unneces- 
sary to say that these bodies are known. We are acquainted with phosphinic 
and phosphonic acids (R'PO(OH) 2 and R' 2 PO(OH)), and with phosphine 
oxides (R 3 PO), substances which are strictly analogous to sulphonic acids 
(RS0 2 (OH)) and sulphones (R 2 S0 2 ), and which are produced by a similar pro- 
cess, viz., by the oxidation of phosphines, but their preparation from phosphoric 
acid or phosphoric anhydride cannot be accomplished. 

Organic Compounds of the three Elements. — Nitrogen is remarkable for 
the ease with which it combines with carbon partly saturated with other 
elements, and consequently the number of organic compounds containing 
nitrogen is very large. The number of these is increased by the fact that 
nitrogen easily combines not only with hydrocarbon radicals, but also with 
radicals containing carbon, hydrogen, and oxygen. Thus the amides are among 
the most numerous of the organic compounds of nitrogen. 

Compounds of sulphur and hydrocarbons are readily obtained, and the mer- 
captans (compounds which may be considered as analogous to primary or 
secondary amines) are also numerous. But compounds of sulphur with 
oxidised organic radicals are scarce. However, we know of thi-acetic acid 
((CH 3 — CO)SH) and sulphide of acetyle ((CH 3 — CO) 2 S), which may be con- 
sidered as analogous to primary (or secondary) and tertiary amides respectively. 

Primary, secondary, and tertiary phosphines are known, and are analogous 
in composition and in many of their properties to amines, but the author is not 



PROFESSOR LETTS ON PHOSPHORTJS-BETAINES. 291 

aware that any phosphorus compound analogous to an amide has been obtained. 
Phosphorus indeed displays but little tendency to combine with oxidised hydro- 
carbon radicals. 

If we compare the phosphines with mercaptans and hydrocarbon sulphides, 
on the one hand, and with the amines, on the other, we find (as might indeed be 
expected) very much the same difference between them as we notice between 
phosphuretted hydrogen, sulphuretted hydrogen, and ammonia. 

Thus compounds of primary phosphines with the hydracids are decomposed 
by water, just as phosphonium iodide is decomposed by water, and the phos- 
phines oxidise with the greatest ease, and even spontaneously. The products of 
their oxidation are analogous to those which the mercaptans and hydrocarbon 
sulphides yield. Thus — 

Pt'SH gives R'S0 2 (OH) 

JRTH 2 » E'PO(OH) 2 
lE' 2 PH „ R' a PO-(OH), 
R' 3 P „ R'gPO , 

as the final products of oxidation. 

The most characteristic property of a mercaptan is the readiness with 
which it exchanges its hydrogen for metals. The author is not aware that 
any attempts have been made to obtain analogous metallic derivatives of 
primary and secondary phosphines, but it is highly probable that such bodies 
may exist and could be easily obtained. 

The organic compounds of the three elements which best admit of com- 
parison are the tertiary amines and phosphines and the sulphides of hydro- 
carbon radicals. These bodies have been well studied, and all of their most 
important properties are known. Let us compare the properties of (CH 3 ) 3 N 
with those of (CH 3 ) 3 P and (CH 3 ) 2 S. They are all volatile liquids of peculiar 
and characteristic odour, and all possess alkaline properties. These are most 
strongly marked in trimethyl-amine, least so in sulphide of methyl. 

Perhaps the most characteristic property of a tertiary amine is the readiness 
with which it combines with the iodide of a hydrocarbon radical to form the 
iodide of a compound ammonium, the hydrate of which is a very powerful base. 
A tertiary phosphine is perfectly similar in this respect, as it combines with 
great readiness with an iodide of a hydrocarbon radical, and from the product of 
union, salts of the compound phosphonium are easily obtained, analogous in a 
great many respects to those of the compound ammonium. A sulphide of 
a hydrocarbon radical also combines readily with the iodide of a hydro- 
carbon radical. Thus on simply mixing sulphide and iodide of methyl, a 
reaction at once occurs, and so much heat is developed by their combination 



292 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

that it is necessary to cool the vessel containing the mixture in order to prevent 
loss. The resulting sulphine iodide is very similar to the iodide of a compound 
ammonium or phosphonium. Its hydrate is a powerful base which absorbs 
carbonic anhydride from the air, and precipitates the hydrates of metals from 
solutions of their salts. 

But there is one important particular in which a tertiary amine is utterly 
unlike a tertiary phosphine, or the sulphide of a hydrocarbon radical. A 
tertiary amine is not capable of directly fixing oxygen, nor indeed of yield- 
ing any simple oxidised derivative. But both a tertiary phosphine and the 
sulphide of a hydrocarbon radical are oxidised with ease (indeed the former 
absorbs oxygen from the air with avidity), and simple products of oxidation 
are formed. Trimethyl-phosphine oxidises to (CH 3 ) 3 PO, and sulphide of 
methyl to (CH 3 ) 2 SO and (CH 3 ) 2 S0 2 . 

Dimethyl-sulphone (CH 3 ) 2 S0 2 and oxide of trimethyl-phosphine closely 
resemble each other. They are solid neutral substances which distil without 
decomposition. They resist the action of oxidising agents in a remarkable 
manner, being unchanged by boiling with nitric acid.'" 

Dr Crum Brown and the author found that just as trimethyl-amine and 
trimethyl-phosphine combine with chloracetic acid to give the hydrochlorate 
of betaine, and of phosphorus-betaine respectively, 

.CI XI 

(CH 3 )=N< and (CH 3 ) 3 =P< 

^CH 2 -COOH ^CH 2 -COOH, 

sulphide of methyl combines with bromacetic acid to give the hydrobromate 
of a body which they called thetine, or rather dimethyl-thetine, 

Br 
(CH 3 )=S<f 

^CH 2 -COOH. 

Thetine is, in certain respects, analogous to betaine. Both are deliques- 
cent bodies, possessing a neutral reaction, and only weak alkaline properties, 
and both yield similar salts, which are readily obtained by the action of silver 
salts on their hydrochlorates or hydrobromates. 

But the author has shown that both the base thetine and also its salts, are 
decomposed by heat in a simple and characteristic manner, whilst Bruhl has 
investigated the action of heat on salts of betaine, and has found that they 
decompose in a completely different way. 

These results of Bruhl's, coupled with the author's experiments, led to the 
question, How will the salts of phosphorus betaine behave when heated ? and 

* The author lias seen their vapour pass almost unchanged over a layer of mixed carhonate and 
nitrate of potash, heated to incipient fusion. 



PROFESSOR LETTS ON PHOSPHORTTS-BETAINES. 293 

in what respects do they resemble and differ from corresponding compounds of 
betaine and thetine ? 

Dr Crum Brown and the author had expressed the opinion that phosphorus- 
betaine would probably more closely resemble thetine than betaine ; and as the 
former had been only subjected to a cursory examination, the author determined 
to carefully examine it, and to study its properties. 

Before, however, proceeding to describe his experiments, it is necessary to 
state the results of Hofmann and Meyer's work on phosphorus-betaines. Hof- 
mann was the first to surmise and to prove their existence, and so far as the 
author can ascertain, he and his pupil Meyer are the only chemists who have 
investigated them. 



"o 1 



" Action of Chloracetic Ether on Triethyl-Phosphine* 

" Triethyl-phosphine and chloracetic ether combine with evolution 

of heat, and formation of a brownish liquid of considerable consistency. If 
somewhat large quantities are to be mixed, it is desirable to moderate the 
action by the presence of a volume of anhydrous ether equal to or greater than 
the aggregate bulk of the two liquids. Dissolved in water, separated by filtra- 
tion or distillation from the excess of chloracetic ether employed, and mixed 
with dichloride of platinum, the new chloride furnishes a beautifully crystal- 
lised platinum salt, which after several recrystallisations from boiling water 
has the composition, 

C 10 H 22 PO 2 ,PtCl 2 = [(C 2 H 5 ) 3 ( C ^A)p]ci ; PtCl . 

submitted to the action of oxido of silver, the chloride undergoes the same 
change which was observed in the corresponding nitrogen compound, 

[(CA)^ : )P]Cl + i|>^}0 = [ (C 2 H 5 ) 3 ( ^}]0 + C A}0. 

It is scarcely necessary to point out the perfect analogy of the new phos- 
phuretted compounds with the corresponding bodies in the nitrogen series. 
Whatever view be entertained of the latter must also be taken regarding the 
former. Conceived in the anhydrous condition, the product obtained by the 
action of oxide of silver upon the chloride may be considered as phosphuretted 
glycocoll, with three equivalents of ethyl in the place of three of hydrogen, 

C 8 H 17 P0 2 = C 2 H 2 (C 2 H 5 ) 3 P0 2 . 

" The phosphuretted compound resembles in its properties the substance 

* Hofmann, Proceedings Royal Society, vol. xi. p. 530. 
VOL. XXX. PART I. 2 Z 



294 PROFESSOR LETTS ON PHOSPHORUS-BETA1NES. 

derived from triethyl-amine. The aqueous solution when evaporated in vacuo 
solidifies into a radiated crystalline mass. I have been satisfied to fix the 
composition of this body by the analysis of the well crystallised platinum salt 
which was found to contain 

C 8 H 18 P0 2 PtCl 3 = [(C 2 H 6 ) S ( C ^)P]C1, PtCl 2 . 

and by that of the iodide. The latter was formed by precipitating the platinum 
salt by sulphuretted hydrogen, decomposing the chloride formed in this manner 
by oxide of silver, and dissolving the triethylated compound in hydriodic acid. 
The solution was evaporated to dryness, the residue washed with absolute 
alcohol, and recrystallised from the same liquid. This iodide is more soluble 
and less beautiful than the corresponding compound in the nitrogen series. 
Analysis showed, however, that it has an analogous composition, viz., 

C 16 H 35 P 2 J = [(C 2 H 5 ) 3 ( C ^ H A)P]I.C 8 H 17 P0 2 



" The Betaine of the Phosphorus Series. # 

" In general the study of phosphorised organic compounds has been preceded 
by the knowledge of the corresponding members of the nitrogen series. There 
are, however, some cases known in which the phosphorised bodies have been 
investigated before the analogous nitrogen compounds. Among these are the 
compounds derived from giycocoll by replacement of hydrogen by alcohol 
radicals. 

" During his great research on phosphorus bases, Professor Hofmann also 
studied the action of monochloracetic acid on triethyl-phosphine, and obtained 
in this reaction the chloride of a base which possesses the composition, 

(C 2 H 5 ) 3 (C 2 H 3 2 )P,C1 . 

" When he removed the chlorine from this body by oxide of silver, there was 
not produced as might have been expected the hydroxyl compound, 

(C 2 H 5 ) 3 (C 2 H 3 2 )P(OH) , 
but by the splitting off of hydrochloric acid, 

(C 2 H 5 ) 3 (C 2 H 2 2 )P = C 2 H 2 (C 2 H 5 ) 3 P0 2 , 
which may be regarded as triethylated giycocoll, the nitrogen of which is 

* Meyer, Ber. d. deutsch. cheru. Ges. iv. 



PROFESSOR LETTS ON PHOSPHORUS-BETALNES. 295 

replaced by phosphorus. At the same time, Professor Hofmann also examined 
the action of triethyl-amine on chloracetic acid, and found, as was to be expected, 
that a corresponding nitrogen body, the triethylated glycocoll, was produced, 

C 2 H 2 (C 2 H 5 ) 3 N0 2 . 

" This last compound received an increased interest when the homologous 
compound in the methyl series, the trimethylated glycocoll, was met with under 
very remarkable conditions. 

" Liebreich showed that the chloride of neurine (a body formed as a decom- 
position product of protagon), and which so far as its composition is concerned 
may be regarded as oxethyl-trimethyl-ammonium chloride, 

(CH 3 ) 3 (C 2 H 5 0)NC1. 

is converted by the action of oxidising agents by replacement in the ordinary 
manner of two atoms of hydrogen by one atom of oxyen into the chloride, 

(CH 3 ) 3 (C 2 H 3 2 )NC1 , 

which, exactly like the homologous ethylated body, by dechlorination loses 
hydrochloric acid, and is converted into the body, 

(CH 3 ) 3 (C 2 H 3 2 )N = C 2 H 2 (CH 3 ) 3 N0 2 . 

that is to say, passes into trimethylated glycocoll. 

" Liebreich obtained the same body, which from its mode of production 
from neurine, may be called oxy-neurine, by a reaction analogous to that which 
Hofmann has given, by acting on trichloracetic acid with trimethyl-amine. 
Some time previously Scheibler, during his researches on the chemical com 
position of the sugar beet, obtained from its juice a splendid crystallised base, 
for which he proposed the name Betaine (from Beta vulgaris). Later 
researches carried out by Scheibler showed that betaine is actually identical 
with the very base obtained from neurine, i.e., with oxy-neurine or trimethylated 
glycocoll. 

" It remained to perfect this group of bodies by the study of the methylated 
phosphorus base. The preparation of the trimethyl-phosphine required for this 
reaction was attended with difficulties, so long as it had to be obtained by the 
former troublesome methods. 

" The reaction discovered by Hofmann and Cahours between trichloride of 
phosphorus and zinc-methyl leaves, it is true, nothing to be desired in sharp- 
ness, but unfortunately all the methods hitherto given for the preparation of zinc 
methyl are in the highest degree uncertain, and give under the most favourable 
circumstances only a very limited yield. On the other hand, the new process 



296 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

which Hofmann has lately communicated, allows of the preparation of trimethyl- 
phosphine in any quantity and of perfect purity. The trimethyl-phosphine 
employed in the research which follows was prepared exclusively by the action 
of phosphonium iodide on methyl alcohol. 

" Action of Monochloracetic Acid on Trimethyl-Phosphine. 

" If the two bodies are mixed in equimolecular quantities a reaction is 
noticed, even at ordinary temperatures. If the mixture has been heated for five 
to six hours in a sealed tube at 100° C, a product is formed which consists of a 
deliquescent viscous syrup, throughout which crystals are interspersed. This 
mass always contains small quantities of hydrochlorate of trimethyl-phosphine 
but consists for the greater part of a compound in which one molecule of tri- 
methyl-phosphine and one molecule of chloracetic acid are united together. It 
was not difficult to demonstrate by means of experiment the composition fore- 
seen theoretically of this body. The directly-formed chloride, on account of its 
hygroscopic properties, and also, as already observed, from its contamination 
with small quantities of hydrochlorate of trimethyl-phosphine, seemed to be but 
little suited for analysis. The simplest way for examining it was clearly the 
analysis of its platinum salt. 

" On dissolving the crude product in water, and adding to the solution a 
slightly diluted solution of chloride of platinum, the platinum salt is precipitated 
immediately as an orange-yellow crystalline mass. This is easily soluble with- 
out decomposition in boiling water, and separates on cooling from this solution 
in splendid crystals of rhombic form. The carbon and platinum determinations 
in this salt showed that it is composed according to the formula, 

C 10 H 24 P 2 O 4 Cl 2 PtCl 4 = 2[C 2 H 2 (CH 3 ) 3 P0 2 ,HCl],PtCl 4 , 

" On treating the aqueous solution of the platinum salt with sulphuretted 
hydrogen the pure chloride is obtained. I concentrated the solution at first on 
the water bath, and then allowed it to stand for some time under the receiver 
of the air-pump : in this manner the chloride is obtained as a crystalline deli- 
quescent mass. On treating the solution with chloride of gold, a beautiful gold 
salt is obtained, which crystallises in long yellow needles, easily soluble in 
water. In order to obtain the free base corresponding with the salts just 
described, the chloride was converted into sulphate by means of sulphuric 
acid. This was treated with baryta, and the excess of baryta removed by 
means of a current of carbonic acid. The solution thus obtained gradually 
solidified in vacuo to a splendid radiating crystalline mass. The solution of 
the base does not affect vegetable colouring matters. If hydrochloric acid is 
added to it, the original chloride is reformed, which was identified by pre- 
paration and analysis of the platinum salt. 



PROFESSOR LETTS ON PHOSPHORUS-BETALNES. 297 

" The base forms with hydriodic and nitric acids well crystallised salts. The 
iodide is easily obtained by dissolving the free base in hydriodic acid, evaporat- 
ing the solution to dryness, washing the dry substance with a little absolute 
alcohol to remove free iodine, and then crystallising the decolorised salt from 
hot alcohol. 

" The iodide is easily soluble in water, and crystallises in beautiful leaflets. 
An estimation of iodine shows that the composition of the salt is, 

C 2 H 2 (CH 3 ) 3 P0 2 .HI. 

"Thus the normal iodide had been obtained, a fact which is somewhat 
remarkable, as, according to Hofmann's researches on the corresponding ethyl 
compound, both of the phosphorus and nitrogen series, a molecule of the base 
is found associated with the iodide. 

" The nitrate is very soluble in water ; the solution crystallised, but less 
easily than that of the iodide. I have not analysed the salt. 

" This also was the case with the free base, which is so hygroscopic that the 
analysis could only have been performed with difficulty. 

" But in the face of so many analogies, it cannot be doubted that in this case 
also, by the action of oxide of silver on the chloride, an exchange of chlorine 
for hydroxyl does not occur, but in its stead a separation of hydrochloric acid, 
in consequence of which the compound would be, 

(CH 3 ) 3 C 2 H 2 2 P = C 2 H 2 (CH 3 ) 3 P0 2 , 

i.e., trimethylated phosphorised glycocoll, or the betaine of the phosphorus 
series." 

The Materials necessary for the Research* 

The materials necessary for the research which the author determined to 
undertake were chloracetic and bromacetic acids, and trimethyl-phosphine 
or triethyl-phosphine. The former were purchased from Messrs Kahlbaum 
& Co. of Berlin, whose preparations the author has always found may be 
relied on. But the trimethyl- and triethyl-phosphine are substances not 
readily purchased, and it was considered better for many reasons to prepare 
them. 

No difficulties were expected in accomplishing this, as Hofmann has 
recently published a method which is stated by him to give excellent results, 

* Owing to the expensive nature of the materials necessary for these experiments the author 
applied in 1879 (when they were commenced) for a sum of money from the Government Research 
Fund, which was granted him. He takes this opportunity to acknowledge the assistance thus received, 
without which he would probably have abandoned the research long before its conclusion. 



298 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

and to be capable of yielding the bodies in any quantity that may be desired. 
This method consists in heating methyl or ethyl alcohol with iodide of phos- 
phonium for some hours in a sealed tube at 180° C, when a mixture of 
hydriodate of triethyl-phosphine and iodide of tetrethyl-phosphonium results, 
from which caustic potash separates triethyl-phosphine in the pure state. 

The author has been unsuccessful in his attempts to prepare trimethyl- and 
triethyl-phosphine in quantity by this process, although he has repeated the 
experiment between thirty and forty times with every precaution. The sealed 
tubes almost always exploded, and only in two or three cases did this not 
occur. He was therefore compelled to abandon this method, and to resort to 
the earlier process for preparing a tertiary phosphine. This was discovered by 
Hofmann and Cahours,* and consists in treating a zinc ether with terchloride 
of phosphorus. The method is at least certain, although tedious and trouble- 
some ; and as zinc methyl is difficult to obtain on a large scale, it was necessary 
for the author to confine his experiments to the ethyl series. 

As he has made very large quantities of triethyl-phosphine by this process, 
and his experience may be of use to others who may have occasion to prepare 
it, he thinks it better to describe the exact method of procedure which he 
adopted. 

Zinc- Ethyl. — This, was prepared by means of the zinc copper couple, which 
Gladstone and TiiiBEt have shown to give very good results on the small scale. 
The author has made a very large number of experiments with this method, 
and always with complete success. The process is simple and easily carried 
out, and the yield of zinc-ethyl is very good. The author can strongly 
recommend it for the preparation of large quantities of that substance. 

The zinc for the couple was always prepared by pouring the molten metal 
into an almost red hot iron mortar, stirring and pounding as rapidly as possible. 
With a little practice, it is easy to manipulate almost 16 kilogrammes of zinc in 
a couple of hours, and to obtain it as a very fine powder. 

This fine powder is sifted from the coarser particles by means of a wire- 
gauze sieve — the gauze being of the usual size employed as a support when 
heating beakers, &c. 

The copper was obtained by the reduction of the ordinary powdered oxide 
of commerce in a stream of hydrogen. It was sifted through the same sieve as 
was employed for the zinc. 

The couple was prepared as Gladstone and Tribe recommend. 

One part of powdered copper and nine parts of zinc powder are placed in a 
flask and heated over a large Bunsen's burner with constant shaking until the 
particles begin to accumulate in small lumps. Great care is necessary to obtain 

* Trans. Roy. Soc. Lond., 1857. 

t Gladstone and Thibe, Journ. Chem. Soc, 1879. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 299 

an active couple, and only practice can insure success. The slightest overheat- 
ing causes the metals to agglomerate into a solid semi-fused mass, which is 
quite inactive. 

The iodide of ethyl was prepared by the usual process, washed thoroughly 
with water, distilled, dried with phosphoric anhydride, and rectified. 

The apparatus employed in the preparation of the zinc-ethyl is shown in 
Plate XVIII. figs. 1 and 2. The iodide of ethyl having been added to the 
couple placed in the flask A, the latter is heated in a water bath B until the 
action is at an end (i.e., until the iodide of ethyl ceases to distil). The 
condenser C is then shifted from the position shown in fig. 1 to the position 
shown in fig. 2, and is connected with the balloon D which contains dry ether, 
and is provided with two tubes and a tap funnel.* The flask A containing the 
ethyl-iodide of zinc is transferred to a bath of melted paraffin B (fig. 2). A 
stream of carbonic acid is passed through the apparatus, and the paraffin 
heated so long as zinc ethyl distils. 

To arrive at the weight of the zinc ethyl produced, the flask containing the 
dry ether (its two tubes stopped with indiarubber and glass rods) is weighed 
before and after the distillation of the zinc ethyl. This flask is then connected 
with the condenser, which in its turn is connected by a cork and bent tube with 
a large tap funnel E (fig. 3), the cork in the tap funnel being provided with a 
small exit tube. A stream of carbonic acid is then passed through the ap- 
paratus, and when the latter is filled with it, the current is stopped by a pinch- 
cock G. The calculated weight of terchloride of phosphorus is now placed in 
the tap funnel F, and the flask D is placed in a water bath, through which a 
stream of cold water is circulating. The tap of the funnel is then opened, and 
the terchloride run in very slowly. The action is violent — the ether boils (in 
spite of the cold water surrounding the flask), and flows over into the tap 
funnel E. When all the terchloride has been added, the water bath is heated, 
and the remainder of the ether distilled off. The tap of the funnel E is now 
opened, and the ether run off. The tap is then closed, and water slowly added 
through F. This usually occasions a violent action, so that it is advisable to 
add the water slowly at first. A large excess of a strong caustic soda solution 
is now added through F, and a layer of the phosphine rises to the surface, a 
white powder (oxychloride of zinc 1) separating also in large quantity. 

The carbonic acid apparatus is disconnected, the water bath removed, and 
a strong current of steam blown through the tube H (fig. 4), which is pushed 
further through the cork, so that its end is almost at the bottom of the flask. 
The phosphine distils over, and forms an oily layer floating on the water, 
which has passed over with it. When no more oily drops distil, the water is 

* The iodide of ethyl ought to be perfectly dry, otherwise a great deal of gas is evolved by its action 
on the couple. The author has found phosphoric anhydride to be the only reliable dehydrating agent. 



300 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 



drawn off and the phosphine collected in a separate vessel. The subjoined 



table gives statistics of the method. 



Preparation of Zinc-Ethyl and Triethyl-Phosphine. 



Weight of Couple. 


Weight of 


Action began 


Action finished 


Zinc Ethyl 


Phosphine 


Iodide of Ethyl. 


in — 




in — 


produced 


produced 


(1) 300 grms. 


300 grms. 


50 rnins. 


15 i 


nius. later. 


95 grms. ) 


75 grms. 


(2) 400 „ 


350 „ 


40 ,, 


10 


>> 


100 „ J 


(3) 300 „ 


300 „ 


15 „ 


20 


}> 


78 „ 


25 „ 


(4) 200 „ 


200 „ 


240 „ 


240 


V 


62 „ ) 
60 „ } 

60 „ J 




(5) 200 „ 


200 „ 


150 „ 


65 


J* 


55 „ 


(6)200 „ 


190 „ 


35 „ 


10 


» 




(7) 200 „ 


200 „ 


120 „ 


120 


■ } 


60 „ | 




(8) 200 „ 


200 „ 


2 days. 


. , , 


63 „ 


(9) 400 „ 


400 » 


30 mins. 


15 


JJ 


115 „ J 




(10) 200 „ 


200 „ 


40 „ 


15 


)> 


... ,j 




(11) 100 „ 


100 „ 


20 „ 


20 


3? 


... j, 




(12) 400 „ 


400 „ 


35 „ 


25 


» 


... j, 


/ (10) and ) 
t (12) gave/ 


(13) 400 „ 


400 „ 


25 „ 




» 


• • • ?> 


47 grms. 


(14) 400 „ 


400 „ 


10 „ 


60 


J> 


115 „ 


40 „ 


(15) 800 „ 


800 „ 


15 „ 


90 


>> 


250 „ 


34 „ 


(16) 300 „ 


330 „ 


10 „ 


120 


>> 


85 „ ) 


53 „ 


(17) 300 „ 


330 „ 


20 „ 


120 


1} 


85 „ f 


(18) 500 „ 


600 „ 


5 „ 


240 


?J 


160 „ 


85 „ 


(19) 350 „ 


470 „ 


7 „ 


90 


» 


117 . J 












1442 


477 



It should be added, that although this method is undoubtedly the best we 
possess at present for preparing triethyl-phosphine, the yield of the latter is not 
very satisfactory, as it amounts to only about 50 per cent, of the calculated 
quantity. Moreover, the crude phosjDhine is by no means pure, and requires 
to be fractionally distilled many times before a product boiling at the right 
temperature is obtained. The higher boiling portions resulting from this 
fractionation consist mainly of oxide of triethyl-phosphine, but the nature of 
the lower boiling fractions (which amount to a considerable quantity) the author 
has not at present been able to ascertain. 

It ought also to be mentioned that triethyl-phosphine is a very disagreeable 
substance to work with for any length of time, as the constant inhalation of the 
small quantities of its vapour produces (as Hofmann has remarked) sleepless- 
ness, which may continue for a considerable time. 

Although triethyl-phosphine oxidises at ordinary temperatures, it does not 
do so to any considerable extent so long as it is not heated. The author has 
always kept it in ordinary stoppered bottles, and has not adopted any special 
precautions while working with it. 

The author's first experiments were made on the action of triethyl-phosphine 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 



301 



on bromacetic acid, as he had employed the latter reagent in his experiments on 
the thetines. Instead, however, of obtaining the hydrobromate of triethyl-phos- 
phorus betaine by this reaction, other bodies were formed, the investigation of 
which was extremely troublesome, and lasted over a considerable time. 

It was only when he substituted chloracetic for bromacetic acid that he 
obtained betaine compounds, and then the investigation was much simplified. 

The author thinks it best to describe these latter experiments first, and then 
to consider the nature of the substances formed by the action of bromacetic 
acid on triethyl-phosphine. 

Action of Chloracetic Acid on Triethyl-Phosphine. 

The apparatus employed in these experiments is shown in the figure. It 
consists of a distilling flask D, which can be immersed when necessary in the 
vessel E, containing cold water and shaken by the hand, the long indiarubber 
tube C permitting this, whilst a stream of hydrogen from the generator A is 
passing through the apparatus. 




A — Hydrogen generator. 

B — Sulphuric acid drying bottle. 

C — Long indiarubber tube. 



D — Small distilling flask. 
E — Vessel full of cold water. 
F— Tap funnel. 



47 grms. of chloracetic acid (molecular weight = 94*5) were placed in the 
distilling flask. The cork through which the tube C (connected with the 
hydrogen generator) and the tube of the tap funnel F pass was then fitted in, 
and hydrogen allowed to stream through the apparatus for some time. 

5*9 grms. of triethyl-phosphine were then placed in the tap funnel F, and 
allowed to drop slowly on to the chloracetic acid. 

The latter dissolved with considerable difficulty in the phosphine, and at first 
no action was apparent, but on shaking the mixture well, a dense syrupy liquid 
began to separate, and after some time most of the mixture had assumed this 
form, but a small quantity of a lighter liquid floated on its surface. The mixture 
grew very hot whilst this was occurring, and had to be cooled repeatedly by 

VOL. XXX. PART I. 3A 



302 PROFESSOR LETTS ON PHOSPHOR US-BET AINES. 

immersing the distilling flask containing it in cold water. In about an hour 
and a half's time the whole solidified to a solid crystalline mass, which was 
perfectly white and very hard. 

This was washed several times with dry ether to remove any phosphine or 
chloracetic acid that had not been acted on. As chloroform had been found to 
be very suitable for dissolving organic phosphorised compounds, an attempt was 
made to get the product into solution by its means, but it did not appear to 
dissolve it perceptibly. Alcohol, however, dissolved it with tolerable ease, and 
on cautiously adding dry ether to the warm solution until the mixture became 
turbid, and then allowing it to stand, beautiful glittering crystals separated, 
which at first appeared to be needles, but afterwards grew into rhombohedral 
plates about 2 mm. in length. 

Nearly the whole of the product was thus recrystallised, then dried in the 
desiccator in vacuo, and submitted to analysis. 

Clilorine. — By precipitation with uitrate of silver. 

(1) 0-4485 gave 0-282 AgCl = 0-07025 CI = 15-6 per cent. CI 

(2) 0-6385 „ 0-3995 „ = 0-09952 „ = 15-6 

Carbon and Hydrogen. — By combustion with chromate of lead : the front of the tube con- 
taining a mixture of the chromate and oxide of copper.* 

(1) 1-296 gave'0-7722 H 2 = 0-0858 H = 6-6 per cent. H 
1-296 „ 2-4013 C0 2 = 0-6549 C = 50'5 „ C 

(2) 0-5165 „ 0-3705 H 2 = 0-04116 H = 8-0 „ H 
0-5155 „ 0-9451 C0 2 = 0-25775 C = 50-0 „ C 





Obtained. 




Calculated for 






f 
i. 


ir. 


(C 2 H 6 ) 3 P< CHi! _ COOH 


(C 2 H 5 ) 3 P< CHs! _ C00C 2 H 5 


Chlorine, . 


15-6 


15-6 


. 


16-7 


. 


14-8 


Carbon, 


50-5 


500 


. . 


45-1 


. 


49-9 


Hydrogen, 


6-6 


8-0 


• 


8-5 


• 


9-1 



These results indicated that the product was not a pure substance. 

The action of chloracetic acid on triethyl-phosphine was repeated with larger 
quantities, exactly the same phenomena being observed as before. 

The product, however, was twice recrystallised from alcohol and ether, and 
was obtained in beautiful colourless needles more than half an inch long. 
These were analysed, and were found to have the composition required for 
the product of union of a molecule of chloracetic acid with one of triethyl- 
phosphine. 

* Neither of these combustions can be relied on as the compound in both cases decomposed with 
unexpected rapidity, and the sulphuric acid in the drying tube blackened, showing that the oxidation 
had not been complete. 



PPOFESSOB, LETTS ON PHOSPHOPUS-BETAINES. 303 

Chlorine. — Volumetrically, by Volhaedt's method. 

(1) 0-5905 required 27 - 9 cc. decinormal AgIST0 3 = 167 per cent. CI 

(2) 0-1930 „ 9-7 „ „ „ = 16-8 „ 

(3) 0-4302 „ 20-2 „ „ „ = 16-6 „ 

Carbon and Hydrogen. 

01497 gave 0-1193 H 2 = 0-013255 H = 8-5 per cent. H 
0-1497 „ 0-2485 C0 2 = 0-06777 C = 45-3 „ C 

The composition of the body was further verified by that of its chloro- 
platinate, which will be described presently. Its reactions indicated that it 
was the hydrochlorate of triethyl-phosphorus-betaine, 

CI 
(C 2 H 5 ) 3 =p/ 

x CH 2 -COOH. 

The experiment of preparing the hydrochlorate was repeated again and 
again, the same phenomena being observed in each case. 

If large quantities are to be operated with (22 grms. of the phosphine and the 
equivalent quantity of chloracetic acid were the largest the author ever em- 
ployed), care must be taken to reduce the acid to a fine powder, otherwise 
it will not dissolve in the phosphine. In any case vigorous shaking of the 
mixture must be resorted to, to accomplish the solution of the acid, and also 
to bring the phosphine thoroughly into contact with it. This is very difficult 
when once the oily layer has begun to form, and only very violent shaking will 
insure the whole of the acid being acted on. In a well-conducted experiment 
scarcely a trace of phosphine remains in excess, but if the shaking has not 
been thorough much remains. Care must also be taken to cool the distilling 
flask ; but, on the other hand, if the temperature is kept too low, the reaction 
is not complete. Some hours were always allowed to elapse before the 
product was recrystallisecl. The recrystallisation is easily effected by dis- 
solving the product in a considerable quantity of hot alcohol, and then adding 
ether cautiously with constant stirring. The addition of ether is stopped as 
soon as the mixture becomes permanently turbid ; on setting it aside for some 
time almost the whole of the hydrochlorate separates in beautiful colourless 
needles. The hydrochlorate thus purified is not perceptibly deliquescent, some 
of the crystals remaining for twenty-four hours exposed to the air without 
liquefying. This is surprising, as Meyer (Joe. cit.) found the corresponding 
compound of trimethyl-phosphorus-betaine to be highly deliquescent. 

It has a sour taste, and an acid reaction. Its other properties will be 
considered later. 

Chloroplatinate of Triethyl- Phosphorus- Betaine. — On mixing dilute aqueous 
solutions of chloride of platinum and of the hydrochlorate, no precipitate is 



304 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

produced ; but if concentrated and hot solutions of the two are mixed, the 
chloroplatinate separates on cooling in groups of thick needles of a light orange 
colour. The compound is a very beautiful one, and frequently the crystals 
attain the length of half an inch. 



(1) 0-5763 


lost 0-0143 2-3 pel- cent 


J$P 


0-5763 


gave 


0-1461 Pt =25-3 


Pt 


0-5763 


» 


0-629 AgCl = 27-0 


CI 


(2) 1-0086 


>> 


0-257 Pt - 25-4 


Pt 


1-0086 


?> 


0-27335 CI = 27-1 


CI 

Calculated for 






M 1-. + .-. -I »-. i.,-1 


A 






' ^ 2 j, (C 2 H 5 ) 3 P<CJ j2C()OH j , PtCl 4 , H 2 C 


Water, 


, 


2-3 


2-3 


Platinum, 


, 


25-3 25-4 


25-2 


Chlorine, . 


. 


27-0 27-1 


27-2 



Hydrobromate of Triethyl-Phospliorus-Betaine. — 9 grins, of the pure hydro- 
chlorate were dissolved in water and converted into hydrate by the action of 
oxide of silver. From the filtered solution the small quantity of silver dis- 
solved was precipitated by hydrobromic acid, and to the solution filtered from 
bromide of silver an excess of hydrobromic acid of constant boiling point was 
added. The mixture was then evaporated in the water bath until a syrup 
remained. This was maintained at a gentle heat to drive off excess of hydro- 
bromic acid. When this had been accomplished the syrup (which was slightly 
brown in colour) was left to itself for a short time, and solidified to a radiating 
crystalline mass. It was dissolved in hot alcohol, and ether was then added 
cautiously to the solution until an oily liquid began to precipitate. The 
mixture was now allowed to stand, and soon began to crystallise. 

The crystals which formed consisted of colourless and very thin quadratic 
plates, which were in many cases half an inch across. 

Some of the crystals were dried in vacuo over sulphuric acid, and a 
determination of bromine made by Volhardt's volumetric process. 

0-0880 required 34 cc. centinormal AgN0 3 = 0*0272 Br=30-9 per cent. Br 

Br 
Calculated for (C,H r ) 3 P< 
Obtained. XJH 2 -COOH. 

Bromine, 30-9 311 

The hydrobromate is somewhat deliquescent, but resembles in other 
properties the hydrochlorate. 

In order to be quite certain that it was really a betaine derivative, some of 
it was reconverted into hydrochlorate by the action of oxide of silver and then 
of hydrochloric acid. This solution yielded the characteristic chloroplatinate 
when it was mixed with chloride of platinum. The necessity for proving that 



PROFESSOR LETTS ON PHOSPHOPTTS-BETAINES. 305 

the body in question was a betaine derivative, and had the constitution 
expressed by the formula, 



(C 2 H 5 ) 3 p/ 



^CH 2 — COOH , 

will be apparent from the experiments on the action of bromacetic acid on 
triethyl-phosphine (see p. 321), for the author was at first led by them to be" 
lieve that no hydrobromate of the phosphorised betaine could exist. How- 
ever, the experiments just described are sufficient to establish the constitution 
of the body in question, which was further proved by the manner in which it 
decomposed when heated (see p. 316). 

Hydriodate of Triethyl-Phosphorus-Betaine. — Hofmann, in the paper already 
mentioned, obtained a hemi-hydriodate of the betaine, 



2 (C 2 H 6 ),P 



on treating the free base with hydriodic acid, and Dr Crum Brown and the 
author have shown that dimethyl-thetine, when treated in the same manner, 
yields a similar compound, 





2 (CH 3 ) 2 S 



But Meyer (he. cit.) obtained the normal hydriodate with trimethyl- 
phosphorus-betaine, viz., 

(CH 3 ) 3 P< CH 2- COOH 

and could not succeed in obtaining a body analogous to Hofmann's hemi- 
hydriodate. 

The author repeated Hofmann's experiment (though in a somewhat modi- 
fied way). 

5 grms. of the pure hydrochlorate were converted into hydrate by the 
action of oxide of silver, and to the filtered solution hydriodic acid of constant 
boiling point was added in slight excess over the quantity required for the 
production of the normal hydriodate. The solution was evaporated to dryness 
on the water bath, and when most of the water had been driven off, yielded 
a syrupy liquid which crystallised on cooling. The solid mass was dissolved 
in alcohol and ether added. 

The hydriodate was precipitated in small granular crystals about as 



306 PROFESSOR LETTS ON PHOSPHORUS-BETA INES. 

large as pins' heads. After some time these were collected and washed 
repeatedly with dry ether. The colourless crystals thus obtained were then 
placed in the desiccator, and after some time the iodine which they contained 
was determined volumetrically by Volhardt's method. 

0-3615 required 11*5 cc. decinormal AgN0 3 = 0146051 Ag — 40 - 5 per cent. 
1-165 „ 37-5 „ „ = 0-476250 „ = 40-8 „ 

Obtained. Calculated for (C 2 H 5 ) 3 P<^ 

£ n x CH 2 -COOH. 

Iodine, 40-5 40-8 . 41-8 

Although these results do not exactly agree with the theoretical quantity, 
it must be remembered that the substance was not recrystallised, and was 
very deliquescent. The author considers that they prove the existence of the 
normal hydriodate, 



(C 2 H 5 ) 3 P<( 



CH 2 -COOH, 

but also believes that the hemihydriodate described by Hofmann exists, and 
probably also an analogous compound of the methylated phosphorus-betaine. 

Hydrate of Trieihyl-Phosphorus-Betaine. — 14 grms. of the carefully purified 
hydrochlorate were dissolved in water and mixed with an excess of moist oxide 
of silver. Chloride of silver was at once formed, and the solution grew warm. 
The mixture was thrown on to a cloth filter, and the solution separated from 
the chloride and oxide of silver by squeezing. A few drops of hydrochloric 
acid were then added to precipitate the silver which had passed into solution, 
and the latter was then filtered. The solution was now placed in vacuo over 
sulphuric acid, and after some days yielded a colourless syrup, which eventually 
solidified to a radiating crystalline mass. 

Some of this crystalline mass was exposed for many weeks in the desiccator 
(but not in vacuo), until it appeared to be perfectly dry.* 

It was then placed in a weighed tube, and exposed in vacuo at first over 
sulphuric acid, and afterwards over phosphoric anhydride. 

The vacuum was not perfect during the five or six months that the compound 
was thus exposed, but was renewed from time to time. 

At the end of this period the hydrate had ceased to lose weight. 
Altogether 045925 lost -0405 H 2 = 8-82 per cent. 

* It was reduced to powder as soon as it was partly dry, so that any water enclosed in the crystals 
might evaporate. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES, . 307 

The equation, 

•CH 9 — COOH _ , /CH„ - CO 

(C 2 H 5 ) 3 P^ OH 2 = (C 2 H 5 ) 3 P/^ o V +H 2 0, 

requires a loss of 9*27. 

The compound which had thus been dried was then burnt with oxide of 
copper and chromate of lead, taking the greatest possible care to introduce it 
into the combustion tube in the dry state, a task of some difficulty, owing to its 
strong attraction for moisture. 



"■a 



0-4160 gave 0-3715 H,0 = 0-0413 H =± 9-8 per cent. H 
0-4160 „ 0-8380 C0 2 = 0-2285 C = 54-9 „ C 

Obtained. Calculated for (C 2 H 5 ) 3 /' * / 

Carbon, . 54-9 54-6 

Hydrogen, . 9-8 , , 9*8- 

Now this result is of some importance, as the base thetine has been shown 

to behave in exactly the same manner; that is to say, when dried over sulphuric 

acid, it has the composition of a hydrate, but in vacuo it loses a molecule of 

water. Thus — 

>CH 2 -COOH /CH 2 -CO 

(CH 3 ) 2 S< =(CH 3 ) 2 S/ / + H 2 



^OH \0- 

And not only in this respect do the two bases resemble each other, for in 
their other properties they are closely analogous. Both are highly deliquescent ; 
both crystallise in the same manner, but only when their solutions are highly 
concentrated ; both have a neutral reaction ; and as will be shown presently, 
both behave in a similar manner when heated. 

Sulphate of TviethyUPhospliorus-Betaine. — This compound Was prepared 
by adding sulphate of silver to a solution of the hydrochlorate, filtering from 
chloride of silver, removing dissolved sulphate of silver by means of hydrochloric 
acid, and after filtering the solution evaporating it in vacuo over sulphuric 
acid. After a considerable time the syrupy liquid which remained, when 
most of the water had evaporated, solidified to a highly deliquescent crystal- 
line mass. 

There can be no doubt that this consisted of the sulphate, 

<C 2 H 5 ) 3 P<^- COOH HOOC - C 5>P(C 2 H 5 ) 3 , 



-SO r 
but owing to its deliquescence it was not analysed. 

Ethyl- Chlorate of Triethyl-Phosphorus-Betaine. — Tliis compound was pre- 
pared by Hofmann, but was not obtained by him in the crystalline state (see 
p. 293). 



308 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

Wishing to obtain some of this body for his experiments, the author pro- 
ceeded in the manner described bv Hofmann, but with somewhat different 
results. 

The following experiment will show this : — 

61 grms. of triethyl-phosphine and 5*9* grms. of chloracetic ether (boiling 
point 140°-145°) were allowed to react on each other in tlie apparatus already 
described for preparing the hydrochlorate, and with the same precautions. 

The two liquids simply mixed at first, but on shaking for some time an 
oily layer was precipitated, and the mixture grew very hot.t By continued 
shaking the upper layer (triethyl-phosphine V) gradually disappeared, and even- 
tually the product consisted of a colourless syrupy liquid. In a few minutes 
very beautiful circular crystals began to appear, and soon the contents of the 
flask had completely solidified. After an hour's interval the white crystalline 
mass thus obtained was broken up and treated three times in succession with 
dry ether. Most of it was then thrown on to blotting paper and placed 
in vacuo over sulphuric acid. 

It was considered sufficient to fix the composition of the body by a chlorine 
determination, and by the analysis of the chloroplatinate. 

1-231 required 50-0 c.c. decinormal AgN"0 3 = 0*1775 CI =14*4 per cent. 

CI 



Obtained. Calculated for (C 2 H 5 ) 3 p/ 
Chlorine, . . 14-4 14*8 



CH 2 -COOC 2 H 5 . 



The ethyl-chlorate of triethyl-phosphorus-betaine is very deliquescent, and 
dissolves with ease in alcohol and chloroform. It cannot, however, be recrys- 
tallised from its solution in either of these liquids, even by the addition of 
ether : the ether precipitating oily drops, which refuse to solidify. 

Chloroplatinate of Ethyl-Chlorate of Triethyl-Phosphorus-Betaine. — A solu- 
tion of the ethyl-chlorate yields a copious precipitate of a light orange colour 
with chloride of platinum, which dissolves on boiling the solution, and separates 
on cooling in crystalline scales. The chloroplatinate thus obtained was ana- 
lysed by dissolving it in water, precipitating the platinum by sulphuretted 
hydrogen, and determining the chlorine in the filtered solution volumetrically. 

0-8245 gave 0*2025 Pt = 24-5 per cent. Pt. 

0*8245 „ 0-8643 AgCl = 0-21532 CI = 26-1 per cent. CI . 

/CI 

Obtained. Calculated for 2 




s CH 2 -COOC 2 Hj 
Platinum, 24-5 . . . 24'0 

Chlorine, 26*1 25-9 

* These quantities are cquimolecular. 

t From time to time the mixture was cooled by immersing the distilling flask in water. 



PROFESSOR LETTS ON PHOSPHORTJS-BETAINES. 309 

Action of Oxide of Silver on the Ethyl-Chlorate. — Hofmann states (loc. cit.) 
that the ethyl-chlorate is decomposed by oxide of silver into the base (triethyl- 
phosphorus-betaine) and alcohol, a reaction exactly similar to that which the 
author has observed with the ethyl-bromate of dimethyl-thetine. Thus — 

.Br .OH 

(CH 3 ) 2 S< + AgHO+H 2 = (CH 3 ) 2 S< + C 2 H 5 + AgBr 

N CH 2 COOC 2 H 5 XJH 2 COOH 

/CI /OH 

(C 2 H 5 ) 3 P< + AgHO + H 2 = (C 2 H 5 ) 3 P< + C,H 5 OH + AgCl . 

N CH 2 COOC 2 H 5 x CH 2 COOH 

The author deemed it of interest to repeat Hofmann's experiment. 

On mixing a solution (in water) of the ethyl-chlorate with recently precipi- 
tated oxide of silver, chloride of silver was formed, and a strong smell of acetic 
ether became manifest. The filtered solution was distilled, and the distillate 
was proved to contain alcohol, but it also had a strong odour of acetic ether. 

The residue of the distillation was mixed with chloride of platinum, and 
yielded the characteristic blunt needles of the chloroplatinate of triethyl- 
phosphorus-betaine. These were dried at 110° C. and analysed. 

0-562 gave 01461 Pt=25-9 per cent. Pt. 
0-562 „ 0-1556 Cl=27'7 per cent. CI . 

Obtained. Calculated for 2 { (C 2 H 5 )P<^ C00H j , PtCl 4 . 

Chlorine, 277 277 

Platinum, 25-9 25-7 

The smell, however, of acetic ether was so pronounced that the author felt 
assured that it had indeed been produced in the reaction. Its formation is 
readily intelligible on the assumption that part of the ethyl-hydrate produced in 
the first phase of the reaction does not break up into alcohol and the phos- 
phorus-betaine, but suffers a totally different, but no less simple decomposition. 

The two reactions may be represented thus — 

/OH /0\ 

(1) (C 2 H 5 ) 3 P< •: = (C 2 H 5 ) 3 P<; \ + C 2 H 5 OH. 

X CH 2 -COOC 2 H 5 x CH 2 -CO 

/0:H 

(2) (C 2 H 5 ) 3 P< / = (C 2 H 5 ) 3 PO + HCH 2 -COOC 2 H 5 . 

\CH 2 -COOC 2 H 5 

The author has not proceeded further with the investigation of this reaction, 
as the experiments to be described presently on the behaviour of the compounds 
of triethyl-phosphorus-betaine with caustic potash fully confirm the above 
interpretation of it. 

VOL. XXX. PART I. 3B 



310 PROFESSOR LETTS ON PHOSPHORUS-BETA INES. 

EthyUBromate of Triethyl-Phosphorus-Betaine. — Bromacetic ether acts on 
triethyl-phosphine with even greater energy than chloracetic ether. It is 
necessary to dilute the bromacetic ether with dry ether before adding the 
phosphine, otherwise so much heat is disengaged that the compound is par- 
tially decomposed. 

Each drop of the phosphine occasions a turbidity in the solution, and the 
ether boils from the heat disengaged unless the vessel in which the experiment 
is conducted is placed in cold water. A layer of oily liquid soon forms, and 
after a few minutes this suddenly solidifies to a solid crystalline mass, whilst 
the supernatant ether also deposits abundance of crystalline matter. Owing 
to the extreme deliquescence of this body, its analysis was not attempted. Its 
reactions, however, leave no doubt as to its composition and constitution, 
which are expressed by the formula, 



Br 
(C 2 H 5 ) 3 =P< 



CH 2 -COOC 2 H 5 . 

This is of some importance, for, as will be shown presently, bromacetic acid 
does not, except under special conditions, give a betaine derivative with triethyl- 
phosphine. 

Ethyl-Iodate of Triethyl-Phosphorus-Betaine. — Iodacetic ether and triethyl- 
phosphine react on each other with as much energy as bromacetic ether and 
the phosphine, and dilution with ether is necessary to moderate the action. 
Exactly the same phenomena are observed as in the preparation of the ethyl- 
bromate. On mixing the ethereal solution of the iodacetic ether with the 
phosphine an oily layer is precipitated, which solidifies after a short time. 
The compound was not analysed. 

Action of Heat on the Compounds of Triethyl-Phosphorus-Betaine. 

One of the most interesting questions which presented itself in connection 
with the compounds which have just been described, was the change which 
they would suffsr when submitted to the action of heat. Indeed, the author 
was chiefly induced to study them from a desire to decide this question. 
For he had shown some time ago * that the compounds of dimethyl-thetine 
are decomposed by heat in a very interesting way, and from the analogy of 
these bodies with corresponding compounds of the phosphorus-betaine, he was 
strongly inclined to the belief that the latter would behave in a similar manner 
to the former when heated. 

The interest of the question was considerably heightened by the fact that 

* Letts. These Transactions, vol. xxviii. p. 591. 



PROFESSOR LETTS ON PHOSPHORUS-BET AINES. 311 

Bruhl* had investigated the action of heat on ethlyated betaine, but had 
obtained a totally different class of products from those into which thetine 
is resolved. 

Thus the action of heat on compounds of thetine and of betaine has been 
studied, with the result that they behave differently. 

How will salts of the phosphorised betaine behave when heated ? Will 
they give similar compounds to those which betaine yields, or will they be 
decomposed in the same way as thetine % 

This question the author determined if possible to decide : but before 
describing the experiments which he performed with this object, it appears to 
him to be advisable to give a short resume of Bruhl's experiments with betaine, 
and of his own with thetine. 

Bruhl's experiments were conducted with triethyl betaine (triethyl-amido- 
acetic acid) 

/CH 2 -COOH /CH 9 -C0 

(0 2 H 6 ) 3 N<r , or (C 2 H 5 ) 3 N< / , 

x OH x O / 



and also with its hydrochlorate, 



(c 2 h 6 ) 3 n/ 



CH 2 -COOH 



CI 

The former was heated in a bath of sulphuric acid to 210° C. At this 
temperature it began to froth and to distil, the distillate consisting of a 
colourless oil possessing an ammoniacal odour. The temperature rose towards 
the end of the operation to 230°, and a small quantity of charcoal remained in 
the retort. The distillate consisted essentially of triethylamine and of the 
unchanged betaine. These two bodies were the only ones which Bruhl could 
obtain from the betaine by the action of heat, but the temperature at which 
the distillation is conducted has a considerable influence on their relative 
quantities. At a temperature of 210°-230°, the quantity of triethylamine is 
from one-third to one-fourth that of the betaine taken, whilst from one-half to 
two-thirds of the latter distils unchanged. 

The only salt of the betaine with which Bruhl appears to have made experi- 
ments was the hydrochlorate. Regarding the action of heat on this, he merely 
says, " I have satisfied myself by experiment that even the chloride (hydro- 
chlorate) distils, but with considerable decomposition." These results appear 
to show that betaines and their compounds either dissociate into a triamine, 

yCH 2 -CO 
and the radical C y/ (or the products of its decomposition), or else distil 

unchanged. 

* Annalen der Chem. u, Pharm., vol. clxxvii. p. 214. 



312 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

Very different is the behaviour of the thetines and their compounds when 
heated. 

The author's experiments extended to dimethyl-thetine itself and to its salts 
with oxyacids and with hydracids. It will be sufficient to say that all the 
former appear to be decomposed in a similar manner, but that the latter suffer 
a totally different kind of decomposition. 

The free base and its oxysalts always yield carbonic anhydride, and a salt 
of trimethyl-sulphine. Its haloid salts give, it is true, a trimethyl-sulphine 
compound, but this is accompanied by thio-diglycollic acid and a haloid ether 
of methyl. 

The decomposition which the free base, its sulphate, and its hydrobromate 
suffer will render this clear. 

( /CH 2 -COOH ) 

(1) 2 (CH 3 ) 2 = S^ = j(CH 3 ) 3 SJ 2 C0 3 + H 2 + C0 2 . 

( OH ) 



(2) 



(CH 3 ) 2 = S<^ (CH 3 ) 2 = S-CH 3 , 

S0 4 = ^>S0 4 + 2C0 2 , 

(CH 3 ) 2 = S/ ; (CH 3 ) 2 = S. 

\ntr _ fnnn 



H ^CH, 



and 



.Br 



(CH 3 ) 2 =S< 

\CH -COOH y CH,-COOH 

(3) = S< + CH 3 Br + (CH 3 ) 3 SBr , 

riTT —poott nhtt —nnnw 



(CH 3 ) 2 = S< 



y CH 2 -COOH N CH 2 -COOH 



Br 



Action of Heat on Ethyl- Bromate of Triethyl-Phosphorus-Betaine. 



* 



Several experiments were made on the action of heat on this body (which 
was always prepared by adding triethyl-phosphine to an ethereal solution of 
bromacetic ether). The results were similar, but as the products were not in 
each case completely investigated, it will be more convenient to give a summary 
of the experiments. 

The ethyl-bromate was heated in a distilling flask connected with the 

* The first of the author's experiments on the action of heat on the compounds of triethyl- 
phosphorus-betaine was made with this body. At the time he had only studied the action of bromacetic 
acid and of bromacetic ether on triethyl-phosphine, and consequently was unacquainted with any salts 
of the phosphorised betaine (see pp. 300, 301). The experiments with the ethyl-chlorate and hydro- 
chlorate were made more than a year afterwards. 



PROFESSOR LETTS ON PHOSPHORUS-BET A INES. 313 

apparatus which the author employed for catching any permanent gases in his 
experiments on the action of heat on salts of thetine.* 

It fused below 100° C, and effervesced between 140°-150° : after some time 
it solidified (at about 170°). A very few drops of liquid distilled, but a large 
quantity of gas was evolved, and of this more than one-third consisted of 
carbonic anhydride. The solid residue was recrystallised (1) from alcohol, in 
which it was very soluble, and another crop of crystals (2) was obtained from 
the mother liquor. 

Bromine was determined in each of these. 

(1) 0-6051 gave 0-5345 AgBr = 0-2275 Br = 37-59 per cent. Br. 

(2) a 0-4104 „ 0-3586 „ = 0-1527 „ = 37"2 
(2) b 0-6648 „ 0-580 „ = 0-2468 „ = 3711 

These numbers agree with the bromine calculated for bromide of triethyl- 
methyl-phosphonium. 

Obtained. Calculated for (C 2 H 5 ) 3 (CH 3 )PBr . 

I. II. 

a b 

Bromine, . . 37-6 . 37-2 371 . . . 375. 

Some of the crystallised product which had yielded these numbers was 
dissolved in water and treated with oxide of silver, hydrochloric acid, and 
chloride of platinum in succession, when an orange-coloured precipitate was 
formed. This chloroplatinate was dissolved in boiling water, and separated as 
the solution cooled in very characteristic crystals, consisting of minute octohedra, 
usually truncated at their solid angles. 

Analysis of these showed them to consist of the chloroplatinate of triethyi- 
methyl-phosphonium. 

Platinum. 

(1) 0-600 gave 0-174 Pt = 29-0 per cent. Pt . 

(2) 0-8563 „ 0-2504 „ = 29-2 

Carbon and Hydrogen. 

0-6703 gave 0-3253 H 2 = 0-03614 H = 5*4 per cent. H 
0-6783 „ 0-6035 C0 2 = 0-1646 C = 24*5 per cent. C 

Calculated for 2 {(C 2 H 5 ) 3 (CH 3 )PCl} , PtCl 4 

29-1 

24-8 
5-3 

These numbers show that the solid product of the action of heat on the 

* These Transactions, vol. xxviii. p. 597. 





Obtained. 




I. II. 


Platinum, 


. 290 29-2 


Carbon, 


. — 24-5 


Hydrogen, . 


. — 5-4 



314 



PROFESSOR LETTS ON PHOSPHORTJS-BETAINES. 



ethyl -bromate is bromide of triethyl-methyl-phosphonium. The formation of 
this substance is explained by the equation — 



^CHj-iCOOiCj^ 



,CH 




(C 2 H 5 ) 3 P< = (C 2 H 5 ) 3 p/ + C0 2 +C 2 H 4 , 

>Br x Br 

a decomposition very similar to that which the oxysalts of thetine suffer when 
heated. 

It must be observed, however, that the author did not detect any ethylene 
amongst the gaseous products of the reaction, but on the other hand only 
about one-third of them consisted of carbonic anhydride. 

Action of Heat on Ethyl- Chlorate of Triethyl-Phosphorus-Betaine. — The 

experiments on the action of heat on this 
body were made chiefly with the apparatus 
represented in the figure. Either the pure 
ethyl-chlorate or the product of action of 
chloracetic ether and triethyl-phosphine was 
carefully weighed, and placed in the flask 
A, which was then connected by the tube 
and indiarubber joint E, with the inverted 
burette B, previously completely filled with 
mercury by raising the reservoir C. The oil 
bath D was then heated by the Bunsen's 
burner G. As the ethyl-chlorate decom- 
posed, the gaseous products passed into the 
burette. Care was taken, by lowering the 
reservoir C, to keep the mercury both in it 
and in the burette at the same level. When 
no further evolution of gas occurred, A was 
cooled in water, and the volume of the gas 
which had been generated read off. The 
proportion of carbonic anhydride which it 
contained was determined as follows. A 
small funnel was attached by an india- 
rubber joint to the tap of the burette, and 
a quantity of strong potash solution poured 
into it. The mercury reservoir was then 
lowered, and the tap of the burette turned on for a few moments until suffi- 
cient potash solution had run into it. The tap was then turned off and the 
gas well agitated with the potash solution, and after some time its volume 
read off. The residual gas could also be examined. 





A — A small flask of about 15 cc. capacity. 

B — Inverted burette filled with mercury. 

C — Mercury reservoir connected by indiarubber 

tube with B. 
D— Oil bath. 
E — Indiarubber junction. 
F — Thermometer. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 315 

A good many experiments were made ; the results being similar in each 
case. The ethyl- chlorate fused below 100° C, and began to give off gas at 
about 120°-130°. As the temperature increased, the evolution of gas became 
more rapid, and eventually the substance solidified, after which no more gas 
was produced. The solid product was brownish in colour, apparently due to 
the presence of charred substances. It ought also to be noticed that a very 
small quantity of liquid distilled during the experiment. 

The quantity of gaseous products amounted to about 14-16 cc. for the 
molecular weight of the ethyl-chlorate in milligrammes, i.e. 0241 grms. The 
addition of potash caused an absorption of about one-half this amount. The 
gas which remained burned with a smoky flame, and was absorbed in great 
measure by bromine,* thus showing that ethylene had been formed. 

The solid product was converted into chloroplatinate. This was re- 
crystallised from boiling water, and the crystals had the characteristic form 
of chloroplatinate of triethyl-methyl-phosphonium. Their identity with that 
substance was proved by their analysis. 

0-4325 grm. gave .... 01285 Pt = 297 per cent. 
0-4325 gave 0-5452 AgCl = 0-1348 CI = 30-9 

Obtained. Calculated for 2 {(C 2 H 5 ) 3 (CH 3 )PCl} ,PtCl 4 . 

Platinum, . . 29-7 29-1 

Chlorine, . . 30-9 31-5 

The preceding experiments show that the ethyl-chlorate decomposes in the 
manner represented by the equation, 

/CH 9 -COOC 2 H 5 /CH 3 

(C 2 H 5 ) 3 P< = (C 2 H 5 ) 3 P< + C0 2 + C 2 H 4 . 

x ci xn 

But this equation can only explain the decomposition of part of the ethyl- 
chlorate ; for if the whole of the latter decomposed as it indicates, a mole- 
cular weight of the substance in milligrames ought to yield at least 44 cc. of 
a mixture of ethylene and carbonic anhydride, whereas, roughly speaking, only 
one-third of that amount of gases were evolved. The author has not, however, 
ascertained the nature of the other reactions or reaction which occur. 

That the ethyl- chlorate should decompose in the manner shown by the 
above equation, the author considers to be somewhat remarkable. When he 
first noticed that the ethyl-chlorate yielded carbonic anhydride on heating, he 
expected that the decomposition had occurred thus — 

<CH 2 — COOC 2 H 5 yCH 2 — C 2 H 5 

= (C 2 H 5 ) 3 < ' +C0 2) 

CI N C1 

but such is not the case. 

* This was ascertained by transferring some of the gas to an inverted burette full of water. A little 
bromine was then added, and the mixture shaken. 



316 PROFESSOR LETTS ON PHOSPHORTJS-BETAINES. 

Action of Heat on Hydrochlorate of Triethyl-Phosphorus-Betaine. — The ex- 
periments made with this substance were conducted in the same apparatus 
as was employed for studying the action of heat on the ethyl-chlorate. 

The hydrochlorate had been recrystallised two or three times from alcohol 
and ether, and was perfectly pure. 

It fused at 145°-150°, then violently effervesced, and after a short time 
solidified to a snow-white mass. The gas evolved consisted of pure carbonic 
anhydride, and amounted to almost exactly the quantity calculated from the 
equation,* 

XJILJCOOiH y CH 3 

(C 2 H 5 ) 3 P< = (C 2 H 5 ) 3 P< + C0 2 . 

X C1 X C1 

The solid product gave the characteristic chloroplatinate of triethyl-methyl- 
phosphonium, which was analysed. 

0-5255 grms. gave 01558Pt = 29-4 per cent. 

0-5255 „ gave 0-6735 AgCl= 01666 01 = 317 „ 

Obtained. Calculated for 2 |(C 2 H 5 ) 3 (CH 3 )PCl} PtCl 4 . 

Platinum, . . 291 29-1 

Chlorine, . . 31:7 ..... 3P5 

Action of Heat on the Hydrobr ornate. — The recrystallised hyclrobromate 
obtained from the hydrochlorate (see p. 304) was also submitted to the action 
of heat. It fused, effervesced from the escape of carbonic anhydride, and 
then solidified. The residue yielded the characteristic chloroplatinate of 
triethyl-methyl-phosphonium. 

Action of Heat on the Hydrate. — The change which the hydrate suffers 
when heated is very interesting, and was discovered quite accidentally. Wish- 
ing to concentrate its aqueous solution, the latter was evaporated on the water 
bath. When most of the water had been driven off, and a syrupy liquid 
remained, the author noticed a smell of triethyl-phosphine. Fearing decom- 
position, the heating was stopped, and the syrupy liquid was placed in the 
receiver of the air-pump over sulphuric acid. On exhausting the air, effer- 
vescence occurred, and the syrup solidified. After some time it was removed, 
and the drying completed on a water bath. 

The dried mass effervesced with acids, even the weakest, such as tartaric 
acid. It at once gave an insoluble precipitate of a light orange colour when 
its solution was mixed with chloride of platinum, and its solution precipitated 

* 78-2 cc. were obtained from 0-762 grms. of the hydrochlorate instead of 80 cc. 



PROFESSOR LETTS ON PHOSPHORUS-BET AINES. 317 

carbonate of silver when it was mixed with nitrate of silver. It also had a 
very faint acid reaction. 

Now all these properties are those of a bicarbonate, and there can be little 
doubt that the hydrate is converted by heat into the isomeric bicarbonate of 
triethyl-methyl-phosphonium. 

Thus — 



/ CH 2 -:C00;H CH 3 

(C 2 H 5 ) 3 =P< • = (C 2 H 5 ) 3 =P< 

X)H X)-COOH. 

That the phosphonium salt had been formed was proved not only by the 
characteristic form of its ehloroplatinate, but also by an analysis of the latter. 

0-5484 grm. gave 01610 Pt = 29-4 per cent. 

0-5484 gave 0-698 AgCl = 0-1726 CI = 31-5 

Obtained. Calculated for 2 {(C 2 H 5 ) 3 (CH 3 )PClf , PtCl 4 . 
Platinum, . 294 .... 29-1 
Chlorine, . 31-5 .... 31-5 

Action of Heat on the Sulphate. — The sulphate when heated behaves like the 
other compounds which have been spoken of. It fuses, effervesces from escape 
of carbonic anhydride, and then solidifies. The solid residue was converted 
into chloroplatinate, which crystallised in the characteristic form of the triethyl- 
methyl-phosphonium compound. It was not considered necessary to analyse it. 

The decomposition of the sulphate, there can be no doubt, is expressed by 
the equation, 

y CH 2 -:COO;'H y CH 3 

(C 2 H 5 )eeP<^ • (C 2 H 5 ) 3 =P^ 

S0 4 = S0 4 + 2C0 2 . 

(C 2 H 5 )=P/ (C 2 H 5 ) 3 =P< 

xjh 2 - ;coo:h n ch 3 

The preceding experiments show that the compounds of triethyl-phosphorus 
betaine behave in exactly the same manner when heated, as the oxy- salts of 
dimethyl-thetine, and in this respect are utterly unlike the compounds of the 
true betaine 

Here then we have another of the many examples in which analogous 
compounds of phosphorus and sulphur display similar properties, whilst the 
corresponding nitrogen compounds behave differently. It should be remarked, 
that the author has in vain sought for a compound of phosphorus analogous 
to thio-diglycollic acid, viz., P(CH 2 COOH) 3 . Neither by the action of heat 
on any compound of the phosphorised betaine, nor by other reactions which 
might be expected to give rise to this body, could it be obtained. The author, 

VOL. XXX. PART I. 3C 



018 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

however, proposes to make other experiments with the view to obtaining it, 
although he thinks it very possible it is not capable of existence. 

Amongst the experiments made in this direction, may be mentioned one in 
which alcohol was heated for more than a week in a sealed tube with the 
hydrochlorate of triethyl-phosphorus-betaine at a temperature varying from 
90°-100° C. Now the hydrobromate of dimethyl-thetine when heated with 
alcohol gives thio-diglycollic ether. Thus — 

/ .CIL.-COOin 

2 (CH 3 ) 2 S^ + 2C 2 H G = S(CH 2 COOC 2 H 5 ) 2 + (CH 3 ) 2 S+2CH 3 Br+2H 2 . 

The phosphorised betaine compound was however simply decomposed, even 
at the temperature mentioned, into carbonic anhydride and chloride of triethyl- 
m ethyl-phosphon ium. 

Action of Caustic Potash on Compounds of Triethyl-Phosphorus-Betaine. 

The author was led to these experiments by an observation he had made, 
that the product of action of brom acetic acid on triethyl-phosphine is readily 
acted on by caustic potash, with formation of oxide of triethyl-phosphine. 

The author was aware that bromacetic acid and triethyl-phosphine do not, 
excejDt under special conditions, give a betaine derivative ; the product formed 
by their union being of a different nature. 

It occurred to him that caustic potash might, however, react with a 
compound of the phosphorised betaine so as to give oxide of triethyl-phosphine, 
and he deemed it of importance to decide this point by experiment. 

Action of Potash on the Hydrochlorate. — A preliminary experiment showed 
that an oily layer at once separated when strong potash solution was mixed with 
the hydrochlorate. 

13 gms. of hydrochlorate (once recrystallised) were dissolved in about 25 cc. 
of water, and solid potash added by degrees. The solution grew very hot, and 
developed a faint odour of triethyl-phosphine, which the author believes to 
have been due to impurities present in the hydrochlorate. When 18 grms. of 
potash had been added, the solution separated into two layers, the lower of 
which consisted of an aqueous solution of the salts formed by the reaction. 
The upper layer was of a yellow colour. It was separated in a tap funnel, and 
fractionally distilled. The thermometer rose rapidly, and remained stationary 
within a degree or two of 240° C, during which a colourless distillate passed 
over, which solidified to a crystalline mass on cooling. 

The boiling-point, zinc iodide compound, and other properties of this body, 
at once characterised it as oxide of triethyl-phosphine. 

It should have been mentioned, that before all the potash had been added 



PROFESSOR LETTS ON PHOSPHORUS-BET AINES. 319 

to the hydrochorate, an attempt was made to extract any substances which 
might have been formed and which were soluble in ether (to which some 
alcohol had been added). The oily layer from which the oxide of phosphine 
was obtained was highly charged with ether, alcohol, and a solid salt, which 
remained in the retort after all the oxide of phosphine had volatilised. This 
was dissolved in water, then just accidulated with nitric acid, nitrate of silver 
added, and the boiling solution filtered from the precipitated chloride of 
silver. The filtered solution was just neutralised with carbonate of ammonia 
and then allowed to cool, when a considerable quantity of crystals separated 
having the appearance of acetate of silver, and which a determination of silver 
showed were really that body. 

0-2444 gave 01565 Ag = 64-0 per cent. Ag . 
Calculated for C 2 H 3 2 Ag = 64-6 „ 

Thus caustic potash acts on the hydrochlorate to give oxide of triethyl 
phosphine, together with chloride and acetate of potassium. The reaction is 
expressed by the equation, 



(C 2 H 5 ) 3 P<f 



CH„-COOH 

+ 2KHO = (C 2 H 5 ) 3 PO + KC1 + CH 3 -COOK + H o 0. 
CI 



Action of Caustic Potash on the Hydrate. — A quantity of the base which had 
been dried in vacuo was shaken with a strong solution of potash. It dissolved 
after a short time, the solution grew warm, and an oily liquid rose to the 
surface. This was separated, and consisted of a strong aqueous solution of 
oxide of triethyl-phosphine.* 

The remaining solution from which the oily layer had been separated was 
neutralised with nitric acid, the mixture heated, and nitrate of silver added. 
On cooling, the characteristic crystals of acetate of silver separated. Their 
composition was verified by a determination of the silver which they contained.! 

0-3157 gave 0-2017 Ag = 63-9 per cent. Ag . 

Calculated for CH 3 - COO Ag = 64-6 

Potash behaves then with the hydrate in exactly the same manner as with 
the hydrochlorate, the reaction occurring as follows : — 

y CH 2 -COOH 
(C,H 5 ) 3 =P< +KHO = (C 2 H 5 ) 3 =PO + CH 3 -COOK + H 2 0. 

\OH 

Action of Potash on the Ethyl-Chlorate. — The author has mentioned (p. 309) 
that, whilst investigating the action of oxide of silver on the ethyl- chlorate, he 

* Oxide of triethyl-phosphine appears to be completely insoluble in strong caustic potasli solution, 
f The crystals became discoloured by pbospburetted hydrogen accidentally present in the air of the 
room in which they were dried. The deficiency in silver is probably due to this. 



320 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

noticed on mixing the two substances a very powerful smell of acetic ether, 
which led him to suspect that part at least of the ethyl chlorate had decom- 
posed according to the equation, 

/CH 2 -COOC H 5 
(C 2 H 6 ) 3 =P< + AgOH = AgCl + (C„H fi ) 3 PO + CH 3 -C(K)Ag. 

N C1 

The action of potash on the ethyl- chlorate has confirmed him in this 
suspicion. On shaking some of the ethyl-chlorate with strong caustic potash 
solution an oily layer separated, and at once a very powerful odour of acetate 
of ethyl was developed. 

It was not considered necessary to proceed further with the experiment, as 
the odour of acetic ether is unmistakable, and the production of the oily layer, 
experience had shown, always indicated the phosphine oxide. There cannot 
be the slightest doubt that caustic potash acts on the ethyl-chlorate, converting- 
it entirely into triethyl-phosphine oxide, chloride of potassium, and acetic ether. 

,CH 2 -COOC 2 H 5 

(C,H 5 ) 3 =P< + KOH = KC1 + (C H 5 ) 3 =P=0 + CH 3 -COOC a H 5 . 

N C1 

Nor can any surprise be felt at this reaction, considering the powerful 
affinity of triethyl-phosphine for oxygen. It is indeed remarkable that such a 
body as the hydrate of the phosphorus betaine is capable of existence at all, 
and still more so that it does not split up into acetic acid and the phosphine 
oxide when heated — 

(C 2 H 6 ) 3 =P^§" COOH - (C 2 H 5 ) 3 PO + CH 3 -COOH. 

The author also tried the action of oxidising and reducing agents on the 
hydrochlorate of triethyl-phosphorus-betaine, but without very interesting- 
results. Nitric acid acted readily on the hydrochlorate when the two were 
warmed together, abundance of red fumes escaping. When all action was over 
the nitric acid was distilled off, and a colourless liquid residue remained, which 
suddenly effervesced at 220° C, red fumes escaping. The heating was stopped 
and the residue was dissolved in water, and heated with chloride of platinum, 
when an abundant light orange-coloured precipitate resulted. Analysis 
showed this to consist of chloroplatinate of triethyl-methyl-phosphonium. Part 
then of the hydrochlorate had escaped oxidation, and had simply lost carbonic 
acid. 

In the nitric acid which had distilled off a small quantity of oxide of 
triethyl-phosphine was detected. The author could find no other definite 
products of oxidation, except a minute quantity of an acid substance which 
gave a white precipitate with sulphate of copper. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 321 



Action of Bromacetic Acid on Triethyl-Phosphine. 

A preliminary experiment showed that a very violent action occurs when 
the two bodies are mixed, so violent indeed that the greater part of the 
mixture was blown out of the vessel in which it was made. 

If, however, the two are mixed in the apparatus employed for preparing the 
hydrochlorate of triethyl-phosphorus-betaine (see p. 301) and with similar pre- 
cautions, the reaction is completely under control. 

The bromacetic acid is at first dissolved by the phosphine, and the mixture 
then grows very hot. If the phosphine is added slowly, and the mixture well 
agitated from time to time, a colourless syrupy liquid results, which does not 
solidify on standing. If, on the other hand, the phosphine is added rapidly, 
and the temperature has not been kept down, the product is dark brown in 
colour, and very often solidifies almost completely on standing. The colour- 
less syrupy product also solidifies on cooling if it be heated for a short 
time at 100° C, but it grows brown during the operation. The solidified 
product is extremely deliquescent, liquefying almost immediately when ex- 
posed to the air. It is very soluble in alcohol, but is insoluble in ether. 
The addition of the latter to its alcoholic solution causes the precipitation 
of an oily liquid which refuses to crystallise. It is also soluble in chloro- 
form, and ether often precipitates it from its solution in that liquid in the 
form of small rhombic crystals. It is, however, extremely difficult to recrys- 
tallise it in this way, and the brown colouring matter adheres to the crystals 
most obstinately. 

The properties of the product either before or after recrystallisation are not 
those of a salt of the phosphorised betaine. Thus it yields no crystalline 
compound with chloride of platinum, nor could a crystalline chloroplatinate be 
obtained after its bromine had been replaced by chlorine (by treating its solu- 
tion with oxide of silver, filtering and adding hydrochloric acid). 

It was found that its solution gave with carbonate or acetate of lead 
crystalline compounds, and much time was spent in endeavouring to fix their 
composition. 

On adding carbonate of lead to the aqueous solution of the product, 
effervescence occurs, and if the solution is hot, a crystalline precipitate is soon 
formed. Also on mixing acetate of lead with a solution of the product, sparingly 
soluble crystalline compounds are produced. If the solutions are cold a white 
flocculent precipitate falls, which in tolerably dilute solution dissolves spontane- 
ously. On scratching the sides of the vessel in which the two solutions have 
been mixed, or on warming the mixture, a colourless salt is precipitated in 
needles or plates. If the solutions are boiling two salts are often formed — one 



322 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

crystallising in warty masses, the other in plates and needles. On dissolving 
either of these in boiling water an insoluble residue is left, which appears to be 
bromide of lead. (It melts to a yellow liquid, and does not char when heated). 
The filtered solution deposits needles or plates on cooling, and very little of the 
warty crystals ; and on again recrystallising, the salt is obtained almost free 
from the latter. 

The composition of the lead salt varies, and although a large number of 
specimens were examined no two of them yielded the same numbers. The 
crystalline form was often entirely different, and was altered by recrystallisation 
of the salt. Moreover, a distinct smell of triethyl-phosphine was always noticed 
when carbonate of lead was employed in its preparation. 

The author could arrive at no definite conclusion as to the composition of 
these sparingly soluble lead compounds. He thinks it advisable, however, to 
give the numbers obtained — 





I. 


II. 


III. 


IV. 


V. 


VI. 


Lead, 


49-4 


49-7 


44-6 


44-5 


43-7 


68-2 


Bromine, 


37-2 


36-6 


40-0 


40-0 


... 


26-7 




VII. 


VIII. 


IX. 


X. 


XL 




Lead, 


67-3 


. . . 


. . . 


53-4 


39-6 




Bromine, 


25-2 


45-2 


42-0 


42-5 


43-0 





I. and II. obtained with acetate of lead, and produced from a hot solution. 
III., IV, and V. „ „ „ a cold „ 

VI. and VII. obtained as I. and II. 
VIIL, IX., X., and XL, obtained with carbonate of lead. 









CH 2 - 


-CO 




/3 = C 8 H l7 P0 2 = 


(C* 


H 5 ) 












Lead. 




Bromine 


Calculated for /3 + PbBr, 


. 


. 


381 




29-4 


/3 + 2PbBr 2 






45-5 




352 


/3 + 3PbBr 2 


. 




48-7 




37-6 


/3 + 4PbBr, 






50-4 




38-9 



The preceding results having failed to establish the composition of the pro- 
duct, other reactions were sought for which would decide this point. In con- 
sidering how to attack this problem the question presented itself, is it not 
possible that the action of triethyl-phosphine on bromacetic acid gives rise to 
an isomer of hydrobromate of triethyl-phosphorus-betaine ? Such a pheno- 
menon would not be extraordinary, as chloracetic, bromacetic, and iodacetic 
acid do not always act in the same manner. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 323 

It is quite conceivable that three bodies can exist having the composition of 
hydrobromate of triethyl-phosphorus-betaine. 

The constitution of these three may be represented by the formulae — 

I. II. 

.Br y Br 

(C 2 H 5 )=P< (C 2 H 5 ) 3 =P< 

N CH 2 -COOH, X)OC-CH 3 , 

Hydrobromate of triethyl- Aceto-bromide of 

phosphorus-betaine. triethyl-phosphine. 



(C 2 H 5 ) 3 P=/ 



III. 

H 

OOC-CH. 2 Br 



Bromacetate of triethyl- 
phosphine. 

No. II. would probably give no platinum salt, whereas No. III., if it gave 
any, would give the chloroplatinate of triethyl-phosphine. No. III. would pro- 
bably give no bromide of silver on treating its solution with nitrate of silver. 

It occurred to the author that the action of caustic potash would decide 

between II. and III. For if it reacted with them at all, the reaction would 

probably be as follows : — 

Bi- 
CII.) (C 2 H 5 ) 3 =P< + 2KHO = (C,H 5 ) 3 =P=0 +CH 3 -COOK + KBr + H.,0. 
X)OC-CH 3 

/ H 
(III.) (C 2 H 5 ) 3 =P< + KHO = (C 2 H 5 ) 3 =P + CH 2 Br-COOK + H,0. 

X)OC-CH 2 Br 

With II. potash would react to give oxide of triethyl-phosphine, acetate of 
potassium, bromide of potassium, and water ; whilst III. would give with the 
same reagent triethyl-phosphine and bromacetate of potassium (or glycollate 
and bromide of potassium). 

It was resolved, therefore, to submit the product of action of bromacetic 
acid on triethyl-phosphine to treatment with caustic potash. 

Action of Caustic Potash. — 18 grms. of triethyl-phosphine were dropped 
slowly into 20 grms. of bromacetic acid in the apparatus already mentioned. 
The product was heated to 100° for about twenty minutes ; it became brown, 
and a few bubbles of gas were evolved ; on standing it solidified. It was then 
dissolved in chloroform, and a large excess of dry ether added — sufficient to 
precipitate the product in the crystalline state. The mixture of chloroform 
and ether was poured off from this, and it was then well washed with dry 
ether, and the last traces of ether removed by gentle heating. 



324 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

It was then dissolved in water and the solution warmed. 6 grins, of solid 
caustic potash were added (dissolved in a little water), and the two solutions 
mixed. No separation of triethyl-phosphine occurred. Another 6 grms. of 
potash were then added ; triethyl-phosphine then separated, but so far as could 
be judged it amounted to only 2 or 3 grms. 

The aqueous solution was drawn off from it, and it was found that the 
addition of strong caustic potash solution to this caused the separation of an oily 
liquid which rose to the surface and collected in a layer. 

The mixture was repeatedly extracted with ether (which dissolved the oily 
layer), and the ethereal extract separated by a tap funnel, and fractionally 
distilled. 

As soon as the ether, water, and triethyl-phosphine had passed over, the 
thermometer rose to 239°, and remained stationary at that temperature, whilst 
a colourless liquid passed over, which solidified on cooling. 

The boiling-point of this liquid, as well as its properties, left no doubt as 
to its identity with triethyl-phosphine oxide. 

The potash solution from which it had been extracted with ether, precipi- 
tated, during the extraction, a colourless crystalline salt. To obtain more of 
this, a considerable quantity of alcohol mixed with a little ether was added. 
The insoluble salt was then collected on a filter, and washed repeatedly with 
alcohol. It weighed 9 grms., and consisted entirely of bromide of potassium. 

These experiments indicate that bromacetic acid unites with triethyl- 
phosphine to give both the isomers, which, for the sake of convenience, we may 
call II. and III. For although neither acetate nor bromacetate of potassium 
were specially sought for in the product of action (owing to the difficulty of 
separating them from the large excess of caustic potash present), the pro- 
duction of both triethyl-phosphine and the phosphine oxide may be considered 
as almost conclusive evidence of the production of both isomers, and from the 
quantities of these it would appear that II. is formed in far larger quantity 
than III. 

But shortly after these experiments were made, it was found that hydro- 
chlorate of triethyl-phosphorus betaine also reacts with potash to give the 
phosphine oxide, and both chloride and acetate of potassium, the reaction 
occurring according to the equation, 

CI 
(C,H.),r< +2KHO = (C^PO + KCl + CHsCOOK + H.O. 

XJH,-COOH 

(see p. 319 . 

The question therefore arose — is no hydrobromate of triethyl-phosphorus- 
betaine formed when bromacetic acid acts on triethyl-phosphine 1 

The hydrobromate was, therefore, prepared from the hydrochlorate (see 



PROFESSOR LETTS ON PHOSPHORUS-BETAIJSTES. 325 

p. 304) and it was found (1) that it readily yielded a sparingly soluble platinum 
salt ; and (2) that it yielded carbonic acid on heating (see p. 316). 

Now it has been already mentioned that no sparingly soluble platinum salt 
could be obtained from the product of action of bromacetic acid or triethyl- 
phosphine, and it had also been found that this product yields only a very 
small quantity of carbonic acid on heating (see p. 328), both of which results 
are against the supposition that any of the true hydrobromate is formed. 

Fresh experiments were, however, necessary to decide this point. 3*5 
grms. of carefully dried and purified bromacetic acid were dissolved in about 
20 cc. of perfectly pure and dry ether. 3 grms. of triethyl-phosphine were 
dissolved in about the same quantity of ether, and the two solutions were 
simply mixed, without any special precautions. The flask in which the mixture 
was made was then corked and placed in cold water : oily drops precipitated. 
The flask was vigorously shaken from time to time, and was then left to 
itself in the cold water. The contents began to crystallise in a short time, 
and soon solidified to a solid mass. After a few hours this was broken up 
and thoroughly extracted with dry ether. It was then placed in vacuo for 
some hours. 

Some of the snow-white product thus obtained was titrated with standard 
nitrate of silver solution, and was found to contain the amount of bromine 
required for the formula C 8 H 18 2 PBr . 

(1) 0-3316 required 13*2 cc. AgN0 3 = 01056 Br =318 per cent. Br . 

(2) 04707 „ 18-5 „ „ =0-1480 „ =314 „ Br. 

Obtained. 



I. ii. Calculated for C 8 H 18 2 PBr . 

Bromine, . 3P8 . 314 . . 31-1 

A portion of the product was treated with oxide of silver, and hydrochloric 
acid was added to the filtered solution. On the addition of chloride of 
platinum to this, a sparingly soluble orange-coloured salt separated exactly 
like the chloroplatinate of triethyl-phosphorus betaine. 

Moreover, on heating some of the product, carbonic acid was given off in 
abundance, no charring occurred, and the residue solidified. On treating the 
latter with oxide of silver, hydrochloric acid, and chloride of platinum in 
succession, the characteristic chloroplatinate of triethyl-methyl-phosphonium 
separated. 

These results then are quite different from those previously obtained, and 
indicate that some at least of the body produced by the action of bromacetic 
acid on triethyl-phosphine is the true hydrobromate of triethyl-phosphorus 
betaine. There was, however, no doubt whatever in the author's mind, from 

VOL. XXX. PART I. 3D 



326 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

the numerous and carefully conducted experiments he had made on the action 
of the two bodies, that under certain conditions none of the true hydrobromate 
is obtained. 

In the experiment just described both the bromacetic acid and the triethyl- 
phosphine were diluted with a large quantity of ether, and the temperature 
was not allowed to rise ; whereas in previous experiments no ether was 
employed as a rule, and the two bodies were allowed to react on each other 
in the pure state. Much heat was developed, and as before stated the product 
of action was frequently heated to 100° C. to cause it to solidify. 

Now it has been shown that the hydrobromate (and other salts) of triethyl- 
phosphorus betaine are decomposed when heated in such a manner that 
carbonic acid escapes, and a salt of triethyl-methyl-phosphonium remains. 

X X 

(C 2 H 5 ) 3 P<f = (C 8 H 6 ) 3 p/* +C0 2 . 

N CH 2 COOH X CH 3 

Whereas the product of action of bromacetic acid on triethyl-phosphine 
yields on heating only a small quantity of carbonic acid, but a large quantity of 
a solid volatile body (see p. 328). It is obvious then that the action of heat is a 
ready method for estimating the amount of hydrobromate of triethyl-phos- 
phorus-betaine present in any specimen of the product of action of bromacetic 
acid on triethyl-phosphine. 

29 grms. of the product just described, and which had been proved to con- 
tain hydrobromate of triethyl-phosphorus-betaine, were heated in an apparatus 
so arranged that any permanently gaseous products could be caught. 

It began to effervesce at 200° C. At 215° C. the effervescence was very 
brisk, and at 230° it suddenly solidified to a pure white product. 192 cc. of 
gas were evolved. 

The solid residue was heated over the naked flame, it fused, boiled, and a 
considerable quantity of a pure white substance passed over at 303° C, which 
solidified in the condenser. Here then is conclusive evidence that the product 
did not consist entirely of the hydrobromate of triethyl-phosphorus-betaine ; had 
it done so no volatile body would have been formed, and 373 cc. of carbonic 
acid would have been produced. In round numbers, only half that quantity of 
gas was evolved, so that at least one-half of the substance consisted of a 
different body from the betaine compound. 

Another experiment was made as follows : — 12 grms. of triethyl-phosphine 
were added rapidly to 14 grms. of bromacetic acid. The mixture was allowed 
to grow very hot, and was cooled only when the phosphine boiled. As soon 
as all action was over, the viscous dark-brown product was divided roughly 
into two parts, one of which was heated in a distilling flask provided with the 
arrangement already described for catching liquid and gaseous products. The 



PROFESSOR LETTS ON" PHOSPHORUS-BETAINES. 327 

heating was performed with a Bunsen's burner, the distilling flask being placed 
on wire gauze. A volatile liquid first passed over, together with about 50 cc. 
of permanent gas. The temperature of the distillate then rose rapidly to 
303° C, and the latter solidified on cooling. No more gas was evolved. 

The other half of the product was dissolved in water, and boiled with slaked 
lime # until the solution was alkaline. Only a trace of triethyl phosphine was 
evolved. The solution was then filtered, mixed with excess of dilute sulphuric 
acid, and the precipitated sulphate of calcium separated from the solution by 
squeezing the mixture on a cloth filter. The dark-brown solution thus obtained 
was distilled until its volume was reduced by about three -fourths. The colour- 
less distillate was saturated with oxide of silver, and the mixture . boiled and 
filtered. 

On cooling abundance of crystalline matter separated, having the appear- 
ance of acetate of silver. It was dried in the desiccator, and a determination 
of silver made. 

0-3202 gave 2057 Ag=643 per cent. Ag. 

Obtained. Calculated for C 2 H 3 2 Ag . 

Silver, .... 64-3 .... 64-6 

Now, in this experiment the triethyl-phosphine and bromacetic acid were 

mixed in the pure state, and the temperature was allowed to rise considerably. 

10 grms. or thereabouts of the product yielded when heated, only 50 cc. of 

gas (presumably carbonic acid) ; whereas, had the product consisted entirely of 

the betaine hyclrobromate, 850 cc. of carbonic acid should have been evolved. 

Therefore only about 5 per cent, of the product consisted of the betaine hydro- 

bromate. Of what did the remaining 95 per cent, consist ? The action of the 

slaked lime may, the author thinks, be considered as proving it to be the 

aceto-bromide of triethyl-phosphine — 

Br 
(C 2 H 5 ) 3 =P<( 

X)OC-CH 3 . 

The lime acting in the same manner as caustic potash, and giving bromide and 
acetate of calcium together with oxide of triethyl-phosphine. 

Br 

2(C 2 H 5 ) 3 =P< + 2Ca(OH) 2 

X)OC - CH 3 

= 2(C 2 H 5 ) 3 =P = O + CaBr 2 + (CH 3 -COO) 2 Ca + H 2 0. 



As before pointed out, any bromacetate of triethyl-phosphine would have been 
detected by the evolution of triethyl-phosphine on the addition of the alkali ; 
whereas in this particular experiment mere traces of that body were given off. 

* Employed instead of caustic potash, on account of its insolubility. 



328 PROFESSOR LETTS ON PHOSPHORUS-BETALNES. 

There is another very powerful argument in support of this view of the 
nature of the product. 

There is no doubt whatever that when it is heated bromide of acetyl is 
evolved (see below). Now, that is exactly what might be expected to occur 
with the aceto-bromide. Thus — 

(C 2 H 5 ) 3 P<J = (C 2 H 6 ) 3 PO + CHs-COBr. 

x o;oc-ch 3 

Action of If eat on the product of action of Bromacetic Acid on 

Triethyl-Phosphine. 

In some of his earlier experiments on the product of action of bromacetic 
acid on triethyl-phosphine, the author had observed that when it is heated a 
crystalline body volatilises. 

This fact seemed to be one of importance, and he therefore determined to 
obtain this crystalline body in quantity, and to examine its properties. 

6 grms. of triethyl-phosphine were mixed in the usual way with 7 grms. of 
bromacetic acid, without diluting the latter with ether. When the action was 
at an end the product was at once submitted to the action of heat. It fused at 
a low temperature ; a few cubic centimetres of gas were evolved, and later a 
small quantity of a pungent fuming liquid distilled. This fuming liquid on re- 
distillation passed over before 100° C. It had the odour of bromide of acetyle, 
and its properties agreed with those of that body. On mixing it with water 
much heat was evolved, and on distilling the mixture (previously diluted with 
a considerable quantity of water) acetic acid passed over, and was identified 
by its silver salt. The residue contained hydrobromic acid. Moreover, on 
mixing some of the fuming liquid with fused acetate of potash, the odour of 
acetic anhydride was at once apparent. There can be no question therefore 
that it consisted mainly of bromide of acetyle. 

After the fuming liquid had passed over the thermometer rose rapidly, 
and a crystalline solid began to appear in the tube used as condenser. The 
distillation was stopped when nothing but a black carbonaceous mass remained 
in the distilling flask. The crystalline solid amounted to about 7 grms. in 
weight. It was melted out of the condensing tube, transferred to a distilling 
flask, and heated. It fused, and at first a little hydrobromic acid was evolved. 
The thermometer then rose to 303° C, and remained stationary,* whilst a 
colourless liquid passed over, solidifying to a white crystalline mass on cooling. 
At the end of the distillation the thermometer stood at 305° C, and about 5 
grms. of the crystalline product were obtained. 

* The condensing tube was changed when the temperature became constant. 



PROFESSOR LETTS ON PHOSPHOPTJS-BETAINES. 329 

It was melted out into a test tube, and three weighed tubes filled with it. 
These were then sealed up, and used for determinations of carbon, hydrogen, 
and bromine. 

Bromine — 

0-6233 gave 04682 AgBr = 019818 Br = 317 per cent. 

Carbon and Hydrogen. (By combustion of the substance with oxide of copper and 
chroniate of lead, a stream of oxygen being passed through the combustion tube at the end of 
the analysis)* — 

(1) 0-45361 gave 0-3184 H 2 = 0-03577 H = 7"8 per cent. H . 
0-4536 „ 0-607 C0 2 = 0-065545 H = 36-5 „ C. 

(2) 0-3329 gave 0-2436 H0 2 = 0-02706 H = 8-1 per cent. H . 
0-3329 „ 0-4515 C0 2 =012313 C = 37'0 „ C. 

In another experiment, conducted in the same manner with 12 grins, of the 
phosphine and 14 grms. of bromacetic acid, the same phenomena were observed- 
17 grms. of crude crystalline product were obtained. This was distilled twice. 
It began to boil at 302°, the temperature was constant at 303°, and the distilla- 
tion was ended at 306°. 

The portion boiling from 302°-304° was at once melted into a test tube, and 
three small tubes were filled for analysis and sealed off. 

Bromine — 

0-7504 grms. gave 0-553 AgBr = 0-234074 Br = 31-2 per cent. 
0-4203 1 „ 0-1320 „ =3P4 „ 

Carbon and Hydrogen — 

0-349 gave 0-2578 H 2 = 0-02864 H = 8-2 per cent. 
0-349 „ 0-4810 „ =0-13117 0=37-6 „ 

Another specimen similarly prepared boiled between 303°-308° C. The 
results of its analysis were as follows . — 

Bromine — Volumetrically. 

(1) 0-296 gave 0-09120 Br=30-9 per cent. Bromine . 

(2) 0-800 „ 0-24880 „ =31-0 

Carbon and Hydrogen — j 

0-4951 gave 0-6584 CO 2 =0-l79564 0=36-3 per cent. Carbon , 
0-4951 „ 03623 „ =0-040255 H=8\L ' „ Hydrogen. 

In these analyses the carbon and bromine agree with the percentages 
required for a product of addition, of one molecule of bromacetic acid and one 
of triethyl-phosphine. 

* This combustion cannot be relied upon, as the substance volatilised with unexpected rapidity, and 
probably some carbonic acid was lost. 

f Volumetrically by Voldhardt's method. 

X This combustion may have given a slight deficiency in carbon, as the substance volatilised very 
rapidly when it was first melted out of the tube. 



330 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

I. II. III. IV. Calculated for C 8 H 18 P0 2 Br 

Carbon, . 36-5 37"0 37-6 36-3 374 

Hydrogen, . 7"8 8-1 8-2 81 7"0 

Bromine, . 31*7 31'2 314 sIhTIsHJ 311 

but the percentage of hydrogen is too high. 

As the bromine was readily precipitated by nitrate of silver, it was con- 
sidered that the body could scarcely be the bromacetate of triethyl-phosphine 

((C 2 H 5 ) 3 P\Qf ) p_nxT -d ), and as the hydrobromate of triethyl-phosphorus- 

betaine ('(CJlo^P^plrT _rtrwyuy nac ^ been shown to give abundance of 

carbonic acid, and to yield a different substance when heated, the new body 
could not be identical with it. 

<TD , 
OOC—CTT 

the aceto-bromide of triethyl-phosphine. It is quite conceivable that it would 
be volatile without decomposition, and it is probable, if not certain, that its 
bromine would be precipitated by nitrate of silver. The evidence appeared to 
be in favour of the identity of this substance with the volatile product in 
question, although the high percentage of hydrogen which the latter contained 
was against this view of its composition. 

The product was very deliquescent, and soluble in alcohol and chloroform, 
but not in ether. It yielded no sparingly soluble compound with chloride of 
platinum neither when alcoholic solutions of the two were mixed nor when it 
was converted into chloride (by action of oxide of silver and hydrochloric acid). 
Attempts were made to determine its vapour density by Victor Meyer's 
method (using vapour of mercury as the source of heat), but without success, 
as it charred. 

It was considered probable that, by acting on it with oxide of silver, its 
nature could be determined. For if its constitution were expressed by the 

/Br 
formula (CkHg^P^ Qr)pi_nTT oxide of silver should give either a correspond- 
ing hydrate, or oxide of triethyl-phosphine and acetate of silver. 



(C 2 H 6 ) 3 P< + 2AgOH = (C 2 H 6 )3PO + AgBr + CH 3 COOAg4-H 2 

\nOf! — f!TT. 



Several experiments were tried on the action of oxide of silver on the 
product. The first of these showed that oxide of triethyl-phosphine is formed. 
The oxide was collected in the pure state; its boiling-point determined, as well 
as other of its characteristic properties. The bromide of silver produced at the 
same time was identified, but no acetate or other soluble salt of silver could be 



PROFESSOR LETTS ON PHOSPHORTJS-BETAINES. 331 

detected. One very carefully conducted experiment may be described to show 
how this was proved. 10 grms. of the product boiling between 304°-306°, were 
dissolved in water and mixed with excess of oxide of silver. Bromide of silver 
was precipitated, but no gas was evolved. The mixture of bromide and oxide 
of silver was then thoroughly squeezed from the solution in a cloth filter, 
suspended in water, and a current of sulphuretted hydrogen passed for some time 
until the mixture was thoroughly saturated. The aqueous solution was then 
filtered off from the sulphide of silver, and was heated in a distilling flask. No 
acetic acid passed over. When hydrobromic acid of constant boiling-point 
began to distil, the residue was heated in a water bath and evaporated to dry- 
ness. A few flakes of crystalline matter (less than 0*5 grm.) remained. 
Neither acetate of silver then, nor any other salt of silver could have been 
precipitated with the bromide except in minute quantity. The aqueous solu- 
tion squeezed from the bromide of silver was heated in a distilling flask con- 
nected with an apparatus for collecting any gas that might be evolved, but none 
came off. Water at first distilled, and later 5-7 grms. of oxide of triethyl- 
phosphine boiling at 240°, and solidifying in the condenser. There remained 
in the distilling flask only a drop or two of a substance which was too small in 
quantity to be investigated. This experiment shows then, that when the 
product is acted on with oxide of silver, only bromide of silver and oxide of 
triethyl-phosphine are produced. 

The results of these experiments are decidedly antagonistic to the view that 
the volatile body consists of aceto-bromide of triethyl-phosphine, and in fact 
may be considered as proving that it is not that substance. They indicate, on 
the other hand, that it consists of a compound of hydrobromic acid with oxide 
of triethyl-phosphine. 

Crafts and Silva* have investigated the action of hydrobromic acid 
on oxide of triethyl-phosphine. By heating the latter with a 64 per cent 
solution of the former to 110° C. they obtained a product which boiled at 
205°-210° C. under a pressure of 2 inches of mercury. This was redistilled 
under a pressure of 1^ inch of mercury, and boiled at 198°-203° C. 

The author subjoins the results of the analyses of these two products, 
together with the mean of the numbers obtained by himself with the volatile 
product boiling at 303° C, and the numbers calculated for a compound of four 
molecules of oxide of triethyl-phosphine with three molecules of hydrobromic 
acid — 



Crafts 


and Silva's 


product boiling 


at — 


Tbe author's 


Calculated for 




205°-210°. 


198°-203°. 




product. 


4[P(C 2 H 6 ) 3 0],3HBr 


Carbon, . . 


. 35-72 


36-18 




36-85 


36-9 


Hydrogen, . 


. 8-03 


8-23 




8-05 


7-7 


Bromine, . 


. 32-17 


31-16 




31-22 


30-8 



* Journal of the Chemical Society, 1871, p. 637. 



332 PROFESSOR LETTS ON PHOSPHORUS-BETA INES. 

Crafts and Silva also passed hydrobromic acid gas into the dry phosphine 
oxide, and distilled the product. It began to boil at 260°, and about half passed 
over at 270-°300° C. A residue was left in the retort at 310°, which began to 
decompose. 

The author considered it advisable to repeat this experiment. 

Action of Hydrobromic Acid on Oxide of Triethyl-Phosphine. 

7--8 grms. of the oxide were fused and a current of hydrobromic acid passed 
through it. The gas was absorbed eagerly, much heat was disengaged, and the 
product was coloured brown. As soon as the hydrobromic acid ceased to be 
absorbed, the product was submitted to distillation. Below 300° a little liquid 
passed over, the thermometer then rose slowly, whilst a colourless liquid passed 
over, which solidified on cooling. It had much the same appearance as the 
volatile product obtained by heating bromacetic acid and triethyl-phosphine, 
but it did not solidify quite so readily as that substance. The thermometer 
was tolerably constant from 320°-325° C, but a good deal of residue remained 
above this temperature. In another experiment the oxide of the phosphine 
was not saturated with hydrobromic acid, but was treated with rather more 
than 30 per cent, of its weight of the gas, which as before was eagerly absorbed. 
On distilling the product thus obtained only a few drops of liquid passed below 
303°. But from this temperature to 308° almost every drop of the product 
passed over, and solidified on cooling to a white solid. 

A determination of the bromine which it contained was made with the 
following: results : — 



■e 



0-3968 required 15-8 cc. decinormal AgN0 3 =31'9 per cent. Br 
04761 „ 19-0 „ „ „ =31-8 

Although these numbers are somewhat higher than those obtained with the 
product of the action of heat on bromacetic acid and triethyl-phosphine, the 
difference is but slight, and very probably it would have been even less had the 
substance been re-distilled. 

The author considers that there can be no doubt as to the nature of the 
volatile body obtained by heating the triethyl-phosphine and bromacetic acid ; it 
is simply a compound of phosphine oxide with hydrobromic acid, or a mixture 
of the two substances, similar to hydrobromic acid, or hydrochloric acid solu- 
tions of constant boiling point. 

Crafts and Silva take the latter view of the nature of the substance 
obtained by them by the action of hydrobromic acid on the phosphine oxide. 
In the memoir already quoted they say, "Hydrobromic, like hydrochloric acid, 
combines with the oxide of triethyl-j3hosphine in the same way that these acids 
combine with water, and it is only under exceptional circumstances that a com- 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 333 

pound with a simple chemical formula is formed." The author, however, is by 
no means convinced of the correctness of this statement, for the numbers 
obtained by them agree very well (as he has shown) with a simple chemical 
formula, and although the latter does not consist of one molcule of the oxide and 
one molecule of the acid, it must be remembered that phosphine oxides combine 
with other bodies frequently in somewhat indefinite molecular proportions, in 
the same manner that silicic acid combines with bases. Further experiments 
are, however, necessary to decide the question. 

The action of heat on the product of union of triethyl-phosphine and 
bromacetic acid cannot be expressed by any simple equation. 

It is, however, probable, from the fact that some bromide of acetyl is 
evolved, that the first action of heat is as follows :— 

Br 
(C 2 H 5 ) 3 =P< = (C 2 H 5 ) 3 PO + CH 3 -COBr. 

X)OC-CH 3 

The phosphine oxide then removes hydrobromic acid from the bromide of 
acetyl, and the residue CH 2 — CO becomes carbonised. 

Action of Bromide of Acetyl on Oxide of Triethy I- Phosphine. 

Whilst the experiments which have just been described were in progress, 
and the author had come to the conclusion that, under certain conditions, 
bromacetic acid and triethyl-phosphine unite to form the aceto-bromide of 
triethyl-phosphine, Dr Ceum Brown suggested that it would be worth while 
to try the action of bromide of acetyl on the oxide of triethyl-phosphine, as 
by that means the same body ought to be formed. 

.Br 
(C 2 H 5 )=P=0 + CH 3 -COBr = (C 2 H 5 ) 3 P< 

X)OC-CH 8 . 

The experiment was accordingly tried. 

The two substances react with energy, and if they are undiluted much 
heat is evolved^ the mixture grows brown, and on cooling solidifies to a 
buttery mass, having exactly the same appearance, and, so far as could be ascer- 
tained, the same properties as the product of action of bromacetic acid on triethyl- 
phosphine. 

On heating, this product behaved exactly like the latter ; hydrobromic acid 
and a small quantity of bromide of acetyl passed over first ; the thermometer * 
then rose to 308° C, and remained stationary at that temperature, whilst a 
colourless liquid distilled, which solidified on cooling, and had the appearance 

* The thermometer employed was different from that used in previous experiments, and the author 
cannot vouch for its accuracy. 

VOL. XXX. PART I. 3 E 



334 PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 

of the product obtained by the action of heat on bromacetic acid and triethyl- 
phosphine, but was rather softer and more buttery. 

Determination of the bromine it contained gave the following numbers : — 

(1) 07472 required 29-8cc decinorrnal AgN0 3 = 31-9 per cent. Bromine . 

(2) 0-909 ,. 36-9 „ „ „ = 32-5 

(3) 0-5665 „ 23-0 „ „ „ = 32-5 

These numbers are somewhat higher than those obtained with the product 
of action of heat on bromacetic acid and triethyl-phosphine, but agree with those 
which Crafts and Silva found in the product of action of hydrobromic acid on 
the phosphine oxide, before it had been re- distilled. 

Although the author feels convinced that all three products have a similar 
composition, he is unable at present to account for the slight differences 
observed in the amount of bromine which they contain. 

The experiment on the action of bromide of acetyl on oxide of triethyl-phos- 
phine may be considered as confirming the view that the author has already 
advanced concerning the nature of the product formed by the action of brom- 
acetic acid on triethyl-phosphine. 

The experiments just described show that the action of bromacetic action on 
triethyl-phosphine varies with the conditions in a very interesting and remark- 
able manner. 

The author thinks that he has proved that, at low temperatures, the two 
substances react so as to produce about equal quantities of hyclrobromate of 
triethyl-phosphorus betaine and aceto-bromide of triethyl-phosphine, or a mix- 
ture of the latter with bromacetate of triethyl-phosphine. 

At intermediate temperatures very little of the hydrobromate is formed, and 
the product consists of the bromacetate and aceto-bromide ; whilst at higher 
temperatures the aceto-bromide is almost the sole product. 

Considering the very powerful affinity of phosphorus for bromine, the trans- 
formation of 

(C 2 H 5 ) 3 =P< 

X)OC-CH 2 Br 

into 

,Br 

(C 2 H 5 )s=P< 

X)OC-CH 3 , 

is readily intelligible, and there can be little doubt that bromacetate of triethyl- 
phosphine is a very unstable body. 



PROFESSOR LETTS ON PHOSPHORUS-BETAINES. 335 

Now, in addition to having a strong affinity for bromine, phosphorus has if 
anything a greater attraction for oxygen, whilst its affinity for carbon is slight, 
so that it is almost surprising that 



(C 2 H 5 ) 3 =P/ 



Br 
CH 2 -COOH 



should be capable of existence at all. And it is certainly a remarkable feature 
in the history of these substances that this body should lose oxygen when 
heated (in the form of carbonic acid). It might rather be expected that it 
would, when heated, be converted into the aceto-bromide. But all attempts 
made in this direction have been unsuccessful. 

In conclusion, I have to express my thanks to my assistant, Mr N. Collie, 
for the assistance he has rendered me during these experiments. 



- 



s. .Roy. Soo.JEdzjz* 7 



Ttz.XKX fiatezmr 







J.BoxtlioIsoiew. Edni 1 " 



( 337 ) 



XII. — On Dust, Fogs, and Clouds. By John Aitken. 

(Read, Part I., December 20, 1880 ; Part II., February 7, 1881.) 

Part I. 

Water is perhaps the most abundant and most universally distributed form 
of matter on the earth. It has to perform more varied functions and more 
important duties than any other kind of matter with which we are acquainted. 
From its close connection with all forms of life, it has been the subject of 
deepest interest in all ages. It is constantly changing from one of its states to 
another. At one time it is solid, now liquid, and then gaseous. These 
changes take place in regular succession, with every return of day and night, 
and every successive season ; and these changes are constantly repeating them- 
selves with every returning cycle. Of these changes, the one which perhaps 
has the greatest interest for us, and which has for long ages been the subject 
of special observation, is the change of water from its vaporous state, to its 
condensation into clouds, and descent as rain. Ever since man first " observed 
the winds " and " regarded the clouds," and discovered that " fair weather 
cometh out of the north," this has been the subject of intensest human interest, 
and at present forms one of the most important parts of the science of 
meteorology, a science in which perhaps more observations have been made 
and recorded than in all the other sciences together. 

In the present paper I intend confining my remarks to this change 
of water from its gaseous or vaporous to its liquid state, with particular 
reference to that change when it takes place in the cloudy condensation 
of our atmosphere. Let us look briefly at the process as it goes on in 
nature. As the heat of the sun increases, and the temperature of the earth 
rises, more and more water becomes evaporated from its surface, and passes 
from its liquid form to its invisible gaseous condition ; and so long as the 
temperature continues to increase, more and more vapour is added to the 
air. This increased amount of vapour in hot air compared to cold air is 
generally explained by saying that hot air dissolves more water than cold air. 
This, however, is not the case. Air has no solvent action whatever on water 
vapour. Water vapour rises into air to the same amount that it would do into 
a vacuum at the same temperature, only it rises into air more slowly than into 
a vacuum, and the amount of vapour which can remain in the air is independent 

VOL. XXX. PART I. 3 F 



338 JOHN AITKEN ON 

of the amount of air present, that is, independent of the pressure of the air, 
and depends only on the temperature. 

After air has become what is called " saturated " with vapour, that is, when 
the vapour tension is that due to the temperature, a momentary condition of 
stability is attained. Suppose the temperature to fall, a change must now take 
place. All the water cannot remain as invisible vapour ; some of it must con- 
dense out into its visible form. It is this condensed water held in mechanical 
suspension in the air to which we give the names of fog, cloud, mist, and rain, 
phenomena having some resemblance to each other, yet possessing marked 
differences. The particles composing a fog, for instance, are so fine they 
scarcely fall through the air, a cloud is a little coarser in the grain, while a 
mist is coarser still in texture, and rain is any of these while falling, whether it 
be a wetting mist or a drenching rain. And the question now comes, Why 
this difference ? Why should the water vapour condense out of the air in one 
case in particles so minute they seem to have no weight, and remain 
suspended in the air, while in another case they are large grained and fall 
rapidly % 

As the key to the answer to this question is given by a very simple experi- 
ment, it will be well for us here to have a clear conception of the conditions of 
that experiment. Here are two large glass receivers, both connected to this 
boiler by means of pipes. If we now allow steam to pass into this receiver, 
which we shall call A, you will see the steam whenever it begins to enter. 
There it comes, rising in a dense cloud, and soon you see the receiver gets filled 
with the condensed vapour, forming a beautiful white foggy cloud, so dense 
that you cannot see through it. Let us now pass some steam into the other 
receiver, which we shall call B. Observe— nay, you may strain your eyes as 
much as you please, you cannot see when the steam begins to enter, and now 
it has been rushing in for some time, and yet you cannot see it. There is not 
the slightest appearance of cloudiness in the receiver, yet it is as full of water 
vapour as the receiver A, which still remains densely packed with fog. 

Now, why this difference in the two cases % Simply this. The receiver A, 
which is so full of fog, was at the beginning of the experiment full of ordinary 
air — the air of this room — while the other receiver B was also full of the air 
of this room, but before entering the receiver it was passed through a filter of 
cotton-wool, and all dust removed from it. The great difference, then, between 
the appearance of these two receivers is due to the dust in the air. Dusty air 
— that is, ordinary air, gives a dense white cloud of condensed vapour. Dust- 
less air gives no fogging whatever. 

But why should there be this difference in the two cases % Why should 
dust have this peculiar action ? or rather, Why does not the water vapour 
condense into its visible form in air free from dust 1 The air is " super. 



DUST, FOGS, AND CLOUDS. 339 

saturated " in both cases, but in the one case it condenses out and forms a 
cloudiness, while in the other it remains in its invisible vaporous form. It 
will be necessary to diverge here a little from our immediate subject, to say a 
few words on the conditions under which water changes from one of its forms 
to another. 

We have what are called the " freezing-point " and the " boiling-point " of 
water. These are, of course, the same as the melting-point and the condensing- 
points of water. Water at 0° C. will freeze if cooled, or melt if heated. It 
will pass into vapour if heated above 100° C, and will pass from vapour to 
liquid if cooled below 100° C, that is, at standard pressure. But something 
more than mere temperature is required to bring about these changes. Before 
the change can take place, a " free surface " must be present, at which the change 
can take place. I may here say that what I mean by a " free surface " is a 
surface at which the water is free to change its condition. For instance, the 
surface of a piece of ice in water is a " free surface " at which the ice may 
change to water, or the water change to ice. Again, a surface of water 
bounded by its own vapour is a " free surface," at which the water may 
vaporize, or vapour condense. What are called the " freezing " and " boiling 
points " of water are the temperatures at which these changes take place at 
such "free surfaces." When there is no "free surface" in the water, we have 
at present no knowledge whatever as to the temperature at which these 
changes will take place. 

It is well known that water may be cooled in the absence of " free surfaces " 
far below the " freezing-point " without becoming solid. Some years ago * I 
showed reason for believing that ice in the absence of " free surfaces " could be 
heated to a temperature above the " freezing-point " without melting. Pro- 
fessor Camelry has quite lately shown this to be possible, and has succeeded in 
raising the temperature of ice to 180° Ct Further, I have shown in the paper 
above referred to, that if water be deprived of all " free surfaces," it may be 
heated in metal vessels while under atmospheric pressure to a temperature far 
above the " boiling-point," when it passes into vapour with explosive violence. 

From this we see that it requires a lower temperature to cause a molecule 
of water to adhere to another molecule of water to form ice, than for a molecule 
of water to adhere to a molecule of ice. Also that it requires a much higher 
temperature to cause a molecule of water surrounded on every side by other 
water molecules to pass into vapour, than for a water molecule bounded on one 
side by a gas or vapour molecule to pass into a state of vapour ; and that a 
necessary condition for water changing its state is the presence of a "free 
surface" or " surfaces," at which the change can take place, if these changes are 

* "Transactions Royal Scottish Society of Arts," 1874-75. 
f " Nature," vol. xxii. p. 435. 



340 JOHN AITKEN ON 

to take place at the " freezing " and " boiling points." At present we do not 
know at what temperatures these changes take place when no " free surfaces " 
are present. Indeed, we are not certain that it is possible for these changes to 
take place at all, save in the presence of a " free surface." 

Returning now to the condensation of the water vapour, we see from the 
experiments given that precisely the same conditions are necessary for the 
condensation of a vapour as for its formation. Molecules of vapour do not 
combine with each other, and form a particle of fog or mist ; but a " free surface " 
must be present for them to condense upon. The vapour accordingly condenses 
on the dust suspended in the air, because the dust particles form " free surfaces " 
at which the condensation can take place at a higher temperature than where 
they are not present. Where there is abundance of dust there is abundance 
of "free surfaces," and the visible condensed vapour forms a dense cloud; but 
where there are no dust particles present there are no "free surfaces," and no 
vapour is condensed into its visible form, but remains in a supersaturated 
vaporous condition till the circulation brings it in contact with the "free 
surfaces " of the sides of the receiver, where it is condensed. 

We see, then, that each fog particle in the experiment was built on a dust 
particle. This indicates an enormous number of dust particles in the air. We 
must not, however, suppose that the particles of that dense fog we saw in the 
receiver A represented all the dust particles in the air experimented on. The 
experiment indicated an extremely foul state of the air indeed, but it does not 
tell the whole truth. Those fog particles only represent a small part of the 
dust particles present. That this is really the case is easily shown in the 
following way : — Let as much steam be blown in as will form a dense fog. 
Now allow this fog to settle, but do not allow any dusty air to enter. After 
the fog has settled blow in more steam. Again you will find a dense fog con- 
densed on the dust which escaped the first condensation. Allow this again to 
settle, and repeat the process a number of times, when you will find, after many 
repetitions, that there is still fog forming. But it will also be noticed that after 
each condensation the fog becomes less and less dense, till at last it ceases to 
appear as fog ; but on closely looking into the receiver the condensed vapour 
will be seen falling as fine rain. When the steam was blown in the first time 
the fog was very fine textured ; each particle was so small it floated easily 
in the air. After each condensation the fog became less dense ; it at the same 
time became more coarse-grained and heavier, and was seen falling slowly. 
Near the end, no fog was visible, and nothing but a fine rain to be seen 
falling. If the air was still further purified, even the rain seemed to cease. 

This experiment may be made in another way. A large globular glass 
flask is provided, having a tight-fitting indiarubber stopper, through which 
pass two pipes. One of these pijDes is connected to an air-pump, and the 



DUST, FOGS, AND CLOUDS. 341 

other terminates in a stop-cock. To the other opening of the stop-cock is 
securely fixed a tube tightly packed with cotton-wool. Some water is placed 
in the flask to moisten the air. If now the stop-cock is closed, and one or two 
strokes are made with the pump, so as to cool the air by expansion, it will be 
noticed that a fog immediately appears in the flask. This fog is fine textured, 
close grained, and will scarcely settle. Now pump out a good deal of the air 
from the flask, and allow air, filtered through the cotton-wool, to enter in its 
place. After the temperature equilibrium is established, again make one 
or two strokes with the pump. The fog again appears, but is now open- 
textured and coarse-grained. Eepeat the process, admitting more and more 
filtered air each time, and it will now be observed that the dense light fog 
which at first appeared gradually gives place to one coarser and coarser in 
texture, till at last no fog appears ; but on looking closely a fine rain, as in the 
previous experiment, will be seen showering down inside the flask. If the 
process is continued still further the rain ceases, there being no more " free 
surfaces " to form nuclei for rain drops. 

These two ways of experimenting, as might be expected, give exactly the 
same result, the conditions being so similar. In one the condensation is pro- 
duced by the cold air mixing with the hot steam ; in the other the " saturated " 
air is cooled by expansion in the flask. These experiments show clearly that 
when there is dust in the air the vapour condenses out in a visible form, but 
when no dust is present it remains in a supersaturated vaporous state. That 
the air, when no dust is present, is really supersaturated, is evident from the 
fact that when the dust particles become few, the fog particles are not only 
few, but are much heavier than when they were numerous, and also by their 
increasing in size as they fall through the air. Each falling particle becomes 
a " free surface," at which the supersaturated vapour can condense and increase 
the size of the drop. Another way of showing the supersaturated condition 
of the air is to allow unfiltered air to enter in place of filtered air. The 
unfiltered air will at once show itself by the vapour condensing on its dust. 
It will be seen rising from the jet into the pure air, falling over and spreading 
itself over the bottom like a fountain of some viscous cloudy fluid. 

It was in the autumn of 1875, when studying the action of "free surfaces " 
in water when changing from one state to another, that I first observed the 
conditions necessary for cloudy condensation. I knew that water could be 
cooled below the freezing-point without freezing. I was almost certain ice 
could be heated above the freezing-point without melting. I had shown that 
water could be heated above the boiling-point, and that the nature of the 
vessel in which it was boiled had no influence on the boiling-point, and all 
that was necessary for cooling the water below the freezing-point and for 
superheating the ice, and the water, was an absence of " free surfaces " at which 



342 JOHN AITKEN ON 

they might change their state. Arrived at this point, the presumption was 
very strong that water vapour could be cooled below the boiling-point for the 
pressure without condensing. It was on looking for some experimental illus- 
tration of the cooling of vapour in air below the temperature corresponding to 
the pressure that I thought that the dust in the air formed " free surfaces " on 
which the vapour condensed and prevented it getting supersaturated. Arrange- 
ments were at once made for passing the air experimented on through a cotton- 
wool filter, and it was then that I first found that air which was free from dust 
gave no cloudy condensation when mixed with steam, and that the super- 
saturated air remained perfectly clear. 

Shortly after this, the investigation had to be abandoned, and all that 
remained of it was a sketch of the apparatus in my notebook, together with a 
description of the experiments made with it, till about the middle of November 
last, when the investigation was continued. The apparatus with which the 
experiments were made before the Society is the same as when used in the first 
experiments. 

The conclusions which may be drawn from these experiments are — 1st, 
that when water vapour condenses in the atmosphere, it always does so on 
some solid nucleus ; 2d, that the dust particles in the air form the nuclei 
on which it condenses ; 3d, if there was no dust in the air there would be no 
fogs, no clouds, no mists, and probably no rain. As we do not at present know 
anything about the temperature of condensation of vapour where there are no 
free surfaces, we cannot tell whether the vapour in a perfectly pure atmosphere 
would ever condense to form rain ; but if it did, the rain would fall from a 
nearly cloudless sky. 

I have said that if there was no dust there would be no fogs, clouds, nor 
mists ; but that is not all the change which would be wrought on the face of 
nature by the absence of dust. When the air got into the condition in which 
rain falls — that is, burdened with supersaturated vapour — it would convert 
everything on the surface of the earth into a condenser, on which it would 
deposit itself. Every blade of grass and every branch of tree would drip with 
moisture deposited by the passing air ; our dresses would become wet and 
dripping, and umbrellas useless ; but our miseries would not end here. The 
insides of our houses would become wet ; the walls and every object in the 
room would run with moisture. 

We have in this fine dust a most beautiful illustration of how the little 
things in this world work great effects in virtue of their numbers. The im- 
portance of the office, and the magnitude of the effects wrought by these less 
than microscopic dust particles, strike one with as great wonder, as the great 
depths and vast areas of rock which, the palaeontologist tells us, is composed 
of the remains of microscopic animals. 



DUST, FOGS, AND CLOUDS. 343 

Let us now look more closely into the action of dust in producing cloudi- 
ness. It is very evident that the results are not always alike. In one case 
the condensed vapour takes the form of a fog, so fine that it easily floats in the 
air and never seems to settle. In another case the cloudiness is coarser 
grained and settles down slowly, and in another case it is a very coarse-grained 
mist which falls quickly (of course I am not here speaking of the coarse 
grainedness produced by a number of small particles combining to form one). 
From the experiments described, it would appear that, when the dust is 
present in great quantities, the condensed vapour forms a fog, because as there 
are a great number of dust nuclei each nucleus only gets a very little vapour, 
and is not made much larger or heavier, so it continues to float in the air. As 
the number of dust nuclei diminish, the amount of vapour condensed on each 
particle increases, their size and weight therefore also increase. So that as 
the density of the cloudiness decreases the size of the particles increases, and 
their tendency to settle down also increases. Fogs will, therefore, only be 
produced when there is abundance of dust nuclei and plenty of vapour. 
There is probably also something due to the composition of the dust particles ; 
some kinds of dust seem to form better nuclei than others. 

We now come to the question of what forms this dust. What is its 
composition % Whence its source ? I have been unable to get any trustworthy 
information as to the chemical composition of the dust. The only analysis 
I have seen is of dust collected in rooms. Now it is evident that as this 
dust has settled down, it will be, so to speak, winnowed dust, and will there- 
fore contain too small a proportion of the finer particles. 

As to where this dust comes from, it is evident it will have many sources. 
Everything in nature which tends to break up matter into minute parts will 
contribute its share. In all probability the spray from the ocean, after it is 
dried and nothing but a fine salt-dust left, is perhaps one of the most important 
sources of cloud-producing dust. It is well known that this form of dust is 
ever present in our atmosphere, and is constantly settling on every object, as 
evidenced by the yellow sodium flame seen when bodies are heated. There is 
also meteoric dust, and volcanic dust and condensed gases. At present, 
however, I wish to confine our attention to the action of heat as a producer 
of atmospheric dust, and more especially in relation to its fog-producing 
power. 

Most of us on entering a darkened room, into which the sun is shining 
through a small opening in the shutters, have observed the very peculiar effect 
of the sun's rays when seen under these conditions, the path of the beam 
of light being distinctly visible, shining like a luminous bar amidst the sur- 
rounding darkness. On closely looking at it, it is seen that this peculiar effect 
is produced by the dust motes floating in the air of the room reflecting the 



344 JOHN AITKEN ON 

light, and becoming visible as they pass through the path of the beam. We 
are struck by the marvellous amount of dust thus revealed ever floating in our 
atmosphere, and which under ordinary conditions of light are not observed. It 
is known that when air containing this dust is highly heated or passed through 
a flame, all these motes are destroyed, and the path of the sun's rays becomes 
invisible. 

Returning now to the question of fogs, one might naturally conclude from 
what we have said that air which had passed over or through a flame or 
through a fire, where the combustion was perfect, ought to be nearly dustless, 
and, therefore, ought not to be a good medium for fogs. Before, however, 
coming to any conclusion on this point, it was deemed necessary to make more 
direct experiments, and we shall presently see that, however natural our con- 
clusion is, it is very far wrong. Heating the air may cause the dust motes to 
become invisible ; but so far as my experiments go, they prove that the 
heating of the air by the flame does not remove the dust, but rather acts in 
the opposite way, and increases the number of the particles. The heat would 
seem to destroy the light-reflecting power of the dust, by breaking up the 
larger motes into smaller ones, and by carbonising or in some way changing 
their colour, and thus make them less light-reflecting. 

Powerful as the sun's rays are as a dust revealer, I feel confident we have 
in the fog-producing power of the air a test far simpler, more powerful and 
delicate, than the most brilliant beam at our disposal. When steam escapes 
into the air it condenses on the dust particles, and thus by simply magnifying 
their size, makes their number evident to the eye. Every fog particle in the air 
was represented by a dust particle before the steam was added, but these were 
invisible to the eye till increased in size by the vapour. This would seem to 
indicate a condition of the atmosphere too impure to be true, yet I think we 
are justified in our conclusion, as it has been shown that when there is no dust 
there is no fogging. In the future, therefore, we will be compelled to look upon 
our " breath " as seen on a cold morning, as evidence of the dusty state of the 
air. And every puff of steam as it escapes into the atmosphere will remind 
us still more powerfully of the same disagreeable fact. If it was not for 
dust we would never see our " breath," nor would wreathes of steam be 
seen floating in the air, nor would our railway stations and tunnels be 
thick with its cloudiness. The only consolation we have is, this fine dust is 
not easily wetted. The air we breathe is not deprived of all its dust in its 
passage through the lungs. The air which we exhale is still active as a fog- 
producer. If, for instance, we inhale the air by the nostrils, and pass it by the 
mouth to the experimental receiver, we find it still full of dust and fog- 
producing. We might have expected, that after passing over so much wetted 
surface, the dust would have been all taken out of the air. This difficulty 






DUST, FOGS, AND CLOUDS. 345 

of wetting the dust in the air may be illustrated by passing air through " washing 
bottles," after which it will still be found to.be full of dust. Further, during 
wet weather, after rain has fallen for a long time, all the dust is not washed 
out of the air. It is still active as a fog-producer, though in a less degree than 
during dry weather. 

I believe that at present some attempts are being made to collect and 
estimate the dust in the air. These observations deal with the weight and 
composition of the dust. I would here suggest that other observations be 
made by this fog-producing power of the air, to get not the weight or compo- 
sition of the dust, but the relative multitude of the dust-specks in it at different 
times. There seems a possibility of there being some relation between dust 
and certain questions of climate, rainfall, &c. 

The composition of the dust will also be of great importance in determining- 
its power as a cloud-producer, as it is evident some kinds of dust will have a 
greater attraction for water vapour than others. Fine sodic chloride dust, for 
instance, we would expect would condense vapour, before it was cooled to the 
saturated point, on account of the great attraction that salt has for water. The 
instrument for these observations might be made to depend, either on the density 
of the fog produced by steam, or on its density when produced by reduction of 
pressure, as in the air-pump experiment. 

Before making any experiment on the fog-producing powers of flames and 
combustion, it was necessary to test the effect of heat on the apparatus to be 
used, so as to be certain the effect was entirely due to the flame and nothing 
due to the heating of the apparatus used in collecting the hot gases. I accord- 
ingly experimented in the following manner : — The cotton-wool filter was 
detached from the experimental receiver, and there was placed between it and 
the receiver a short length of glass tube, so arranged that the air after passing 
through the filter should pass through the tube on its way to the receiver. The 
tube was so arranged that it could easily be taken out to be cleaned, and opened 
for introducing into it any substance the effect of which we might wish to test. 
The receiver was connected to an aspirator, by means of which filtered air was 
drawn into the apparatus. 

The glass tube was first carefully washed with soap and water, and then 
with sulphuric acid, the acid being carefully washed off before the tube was 
put in its place. Air was now drawn through the apparatus, the air being- 
tested from time to time by the admission of steam into the receiver. At 
first the steam gave rise to cloudiness, but as the dust gradually got cleared 
out the clouding become less and less, till at last it disappeared, indicating 
a dustless state of the air in the receiver. After this condition was attained 
the glass tube, through which the filtered air was passing, was heated, 
to get the effect, if any, due to heating glass, and also to make sure that the 

VOL. XXX. PART I. 3 G 



346 JOHN AITKEN ON 

effect produced by any substance placed in the tube was due to that substance 
alone. The result of heating the clean and empty tube was most remarkable, 
and very unexpected. A slight heating was sufficient to give rise to a very 
dense fog, on admission of steam to the receiver. We might have imagined 
that the careful washing the tube received was sufficient to make the glass 
clean. Yet we see it was still so foul that heat drove off sufficient matter in a 
fine state of division as to give rise to a dense fog. The glass tube was now 
highly heated, to see if heat would cleanse it. After cooling it was again 
heated to the same amount as at first. It was now found to be quite inactive. 
No fogging whatever appeared in the receiver. If, however, the tube was again 
highly heated fogging appeared. In testing different substances placed in the 
tube, it was therefore necessary to use only a low degree of heat, so that none 
of the effect might be due to the tube. After each experiment the tube was 
highly heated, to thoroughly cleanse it, before introducing the substance to be 
tested. When this was done, and a lower degree of heat employed, I could 
perfectly trust to the tube being inactive. 

The next experiment was made with a small piece of brass wire placed in 
the testing tube. While it was cold there was of course no fogging, but when 
slightly heated, a dense clouding resulted. A piece of iron wire, and other 
substances, all gave a similar result. The wires were now highly heated in a 
Bunsen flame before being put in the testing tube. On heating they were 
now found to be quite inactive, not the slightest fogging appeared. The high 
temperature had acted on them as it acted on the glass, and destroyed their 
dust-producing powers. 

A piece of brass wire was now carefully filed bright, so as to remove all 
uncleanness from it, it was then placed in the experimental tube, care being 
taken that it was not touched with the hands. When heated it only gave 
rise to the faintest cloudiness. These experiments prove that the cloudiness 
was produced by some matter driven off by the heat from the outside of the 
metal. The slight cloudiness produced by the filed wire being due to the 
slight contamination got when being filed. 

The amount of matter which is driven off these wires by heat is extremely 
small, and its result as a fog-producer so great, that this apparatus places in 
our hands a means of detecting in gases quantities of matter so small as 
almost to rival in delicacy the spectroscope. The following experiment will 
give an idea of the marvellous smallness of the amount of matter which may be 
detected in this way. If we take a small piece of fine iron wire, ^j of a grain 
in weight, and place it in the experimental tube, and apply heat, it will give 
rise to a very decided cloudiness. Now take the wire out, and if you so much 
as touch it with your fingers, on again returning it to the tube and heating, 
the fact of your having touched the T ^y of a grain of iron wire will be declared 



DUST, FOGS, AND CLOUDS. 347 

by the fog which forms in the receiver. The effect seemed so great for so 
small a cause, that I repeated the experiment a great number of times, some- 
times putting in the wire and getting the fog, and sometimes going through all 
the motions and changes necessary for, but not putting it in, and getting no 
fog, that I am compelled to come to the conclusion, that the fogging is really 
caused by the contamination due to the touch. 

A great number of different substances were tested in this apparatus, and, 
as might have been expected, all were active fog-producers. Amongst other 
substances tried were different salts. One point noticed was that their 
activity did not depend on their power of evaporating or subliming. Camphor, 
though subliming and evaporating quickly, scarcely ever gave any fog, only a 
heavy coarse-grained fog which settled at once, while amnionic carbonate, 
soclic carbonate, and sodic chloride were very active, indeed the latter salt is 
one of the most active substances I have tried. If we place a crystal of sodic 
chloride j^ grain in weight in the tube, and apply heat, it will continue to 
give off nuclei sufficient to form a dense fog for a long time, without apparently 
losing in size. 

We see from these experiments that when testing the fog-producing power 
of a flame, it will not do to collect the products of combustion and draw them 
into our experimental receiver, as the heat would raise a dust from the surface 
of the collecting tube sufficient to cause a dense fog ; another method of 
experiment was therefore devised. It was, however, necessary before pro- 
ceeding further, to test the effect of the gas to be burned, to see if it was 
active as a fog-producer. Gas from the gas pipes was accordingly passed into 
the experimental receiver, and tested with steam, and found to be perfectly 
inactive. No cloudiness appeared. Any effect then produced by the burning 
gas could not be due to dust carried in by the gas. 

The apparatus was now arranged in the following manner to test the fog-pro- 
ducing powers of the products of combustion from a gas flame : — Two receivers 
were arranged alongside each other, and connected by means of a pipe. 
Gas was led into the first receiver by a pipe terminating a short distance 
inside the receiver in a glass tube, the end of which was drawn to a fine jet 
at which the gas was burned. The receiver used for this purpose was so large 
that the flame could not heat the glass sufficiently to make it active as a fog- 
producer. After the gas was lighted, a current of filtered air was drawn through 
the receiver to supply oxygen for the flame. The products of combustion were 
drawn into the second receiver through the connecting pipe. In this second 
receiver the products of combustion were tested from time to time with steam. 

At first, of course, the air which came would be unfiltered dusty air ; 
but as nothing but filtered air entered, this dusty air ought gradually to give 
place to pure air. It was found, however, that after filtered air had been 



348 JOHN AITKEN ON 

drawn through for a long time, there was not the slightest sign of the air 
becoming purer. To make sure the fogging was due to the flame, the gas was 
turned off, and combustion stopped, while the circulation was kept up. In a 
very short time after this was done, the air showed a marked decrease in 
cloudiness, and after a time became pure. 

This method of testing the effect of combustion does not seem, at first 
sight, the best. The intention was to have, first, circulated the air till perfectly 
pure, and steam gave no cloudiness, and then to light the gas and see the 
effect. The difficulty of working in this way was that I could not light the 
gas without introducing a disturbing element. It was intended to have lit the 
gas by means of an incandescent platinum wire, but on testing the effect of 
the hot wire alone, it was found to make the air active, and powerfully fog- 
producing. By highly heating the wire, it was possible to make it less active 
at lower temperatures, but the temperature produced by igniting the gas would 
again make it active. 

I have great hesitation in coming to any conclusion from this experiment. 
At first sight it would look as if the small flame is very far from being a dust 
destroyer, and is on the contrary a very active producer of it. It will be 
remembered that the flame was fed with filtered air, and the result of the 
combustion of filtered air and dustless gas is an intensely fog-producing atmo- 
sphere, and that the fogging is due to dust cannot, I think, be doubted, as the 
products of combustion, when filtered, give no cloudiness when steam is added. 
Yet the question may be asked, Was the dust produced by the combustion % 
It seems almost possible it might be the result of soda driven off by the heat 
from the glass jet. 

On the 8th and 12th of January this experiment was repeated. The glass 
jet at which the gas was burned being removed, and a platinum one put in its 
place. Platinum was selected because it was thought in the highest degree 
improbable that any nuclei could be driven off the platinum by the heat of 
the gas flame. After the jet was fixed in its place it was highly heated to 
thoroughly cleanse and make it inactive at the lower temperature produced by 
the flame. The gas was lit, and the receiver then put in its place, and the 
supply of filtered air drawn through the apparatus. The result was the same 
as before. Increase of fogging on the gas being lighted, and the fogging con- 
tinued so long as the gas was kept burning, and only stopped when the flame 
was put out. 

There seemed a possibility that the fogging might be due to some residual 
motes still remaining in the receiver getting into the flame and being broken up 
by the heat into a great nupiber of parts. The experiment was accordingly 
varied to meet this. A fine platinum wire, which could be heated by a battery, 
was arranged so that the gas might be lit by it without opening the receiver, 



DUST, FOGS, AND CLOUDS. 349 

the platinum wire being previously highly heated to cleanse it as much as 
possible. The receivers being closed, and the gas not lit, air was drawn 
through the apparatus till the air in the receivers was purified ; and no cloudy 
condensation took place on admitting steam. Contact with the battery was now 
made, and the gas lit. At once a densely fogging atmosphere was produced. 

No doubt part of this fogging was due to nuclei driven off the heated 
platinum wire, but as the wire was previously cleansed, and only heated for a 
short time, and quickly removed from the flame, there would be but little due 
to this cause, and what dust it did give off would be so fine that the heat of 
the flame would not be likely to break it up any further, and it would be 
gradually removed by the circulation, and its place filled with filtered air. It 
was, however, found that though the supply of air was kept up, and the flame 
kept burning for some time, the fogging showed no signs of decreasing. On 
shutting off the gas, the fogging at once began to diminish, and soon cleared 
away, showing that the fogging was due to the products of combustion. 

These experiments seem to indicate that the combustion of clustless gas 
and dustless air do of themselves give rise to condensation nuclei, and do not 
act by simply breaking up larger dust motes into smaller ones. These nuclei 
produced by the combustion of gas must be extremely small, as a very small 
flame so loads a considerable current of dustless air as to cause it to become 
full of a very fine and closely packed form of fog when mixed with steam. 

The question may here be put, Is it really dust which is driven off by the 
heat from the surface of glass, from the brass and iron wires, and from the 
other substances ? It is extremely difficult to get a direct answer to this 
question, but I think that, reasoning from the known conditions necessary for 
the condensation of vapour, it is extremely probable that it really is an ex- 
tremely fine form of solid matter which is produced under these circumstances. 
Further, they have all been put to the test of the cotton-wool filter, and all of 
them have been filtered out and the air made non-cloud-producing. If it was 
some gas or vapour which was produced by the heat, we see no reason why the 
cotton-wool should have kept them so completely back. 

Another set of experiments was now made to test the fog-producing power 
of air and gases from different sources. The air to be tested was introduced 
into the experimental receiver, and steam blown in and mixed with it. Its fog- 
producing power was tested by the density of the cloudiness produced, and also 
by the time the fogging took to settle. It was always found that the air of the 
laboratory when gas was burning gave a denser fog than the air outside, some- 
times two or three times as dense. The products of combustion from a Bunsen 
flame and from a smoky flame were compared. They were found to be about 
equally bad, and both much worse than the air in which they were burned. 
These products were collected by holding the open end of the receiver over the 



350 JOHN AITKEN ON 

flame, taking care not to heat the glass. Products of combustion from a clear 
part and from a smoky part of a fire were tested, and found to be about equally 
foggy, and both much worse than the air of the room. 

From these experiments it would appear that combustion under all condi- 
tions is bad as a fog-producer ; bad, whether the combustion be perfect, as in a 
Bunsen flame and a clear fire, or imperfect as in a smoky flame and smoky fire. 
It is therefore hopeless to expect that by adopting fires having a perfect com- 
bustion, such as the gas ones now so much advocated, we would thereby 
diminish the fogs which at present, under certain conditions, envelop our 
towns, and give rise to so much that is both disagreeable and detrimental. All 
fires, however perfect the combustion, are fog-producers when accompanied by 
certain conditions of moisture and temperature. From this it will be observed 
that it is not the visible dust motes seen in the air that form the nuclei of fog 
and cloud particles, as these may be all destroyed by combustion, and yet the air 
remain fog-producing. No doubt these motes also play their part in the condensa- 
tion, but their number is too small to be of importance. The fog and cloud nuclei 
are a much finer form of dust, are quite invisible, and though ever present in 
enormous quantities in our atmosphere, their effects are almost unobserved. 

A number of experiments have been made by burning and highly heating 
different substances to test their fog-producing powers, and I have found that 
highly heated sodic chloride, as, for instance, when burned in an alcohol flame, 
or salt water spray heated in a Bunsen flame, gives rise to an extremely dense 
fog when tested with steam. But perhaps the most active of all substances I 
have yet tried is burning sulphur. The fog produced when steam has been 
blown into air in which a very little sulphur has been burned is so dense that if 
ever fog was " cut " it might or should be. So dense is it that it is impossible 
to see through a depth of more than 5 centimetres of it. The sulphides when 
burned also give similar results. 

These experiments evidently introduce a new element into the investigation. 
We have here not only to do with the attraction of the different molecules of 
the same kind, but the gaseous molecules in this case have also chemical 
affinities for each other. It is very difficult to understand this marvellous 
fog-producing power of burned sulphur. Sulphur in burning gives rise to 
sulphurous acid. Now from experiment I have made with sulphurous acid 
prepared from sulphite of soda and sulphuric acid, and also from copper and 
sulphuric acid, the sulphurous acid being carefully dried with sulphuric acid, I 
do not find it active as a fog-producer. It gives rise to no fumes, it does not 
increase the fogging of dusty supersaturated air, and produces no fog in filtered 
supersaturated air. 

Sulphuric acid vapour, it is well known, gives rise to dense fumes by com- 
bining with the moisture of the air, and I find, under certain conditions, it also 



DUST, FOGS, AND CLOUDS. 351 

gives rise to a dense fog with steam, but I also find that these fumes and fog 
owe their formation to dust. This is illustrated by the following experiment. 

In a retort was placed a quantity of sulphuric acid. The stopper of the 
retort was removed, and in its place was put a tube connecting the retort with 
a cotton- wool filter. The neck of the retort was connected to a wash-bottle by 
means of a glass tube. An aspirator drew the air out of the wash-bottle, and 
thus kept up a current of air from the filter through the retort to the wash- 
bottle, the air bringing the sulphuric acid vapour along with it. At first, when 
unfiltered air passed, dense fumes filled the retort and wash-bottle, but when 
the filter was introduced the cloudiness gradually disappeared. The absence 
of dust entirely prevented any foggy condensation, even though there were 
chemical affinities. After the experiment had been continued for some time, 
slight fumes began to appear, even when filtered air was passing, but this only 
happened when the acid became very concentrated, and much acid evaporated, 
and the fumes with filtered air were very slight, while unfiltered air gave very 
dense fumes. 

It is not necessary to suppose the want of dust prevented the chemical 
affinities from acting, it only prevented the new compound from condensing 
in cloud form. When the acid was weak its vapour would combine with the 
moisture in the air, but would remain as vapour when there was no dust for it 
to condense upon. But when the acid became highly concentrated, the mole- 
cular strain would be greatly increased on account of the vapour tension being 
greatly in excess of that due to the temperature, and it would then seem to be 
able to condense without the presence of a " free surface." There is, of course, 
the possibility that the filtering of the air was not perfect. I may remark here 
that the fumes of highly concentrated sulphuric acid are found to be an excellent 
fog-producer. If we dip a glass rod in the acid, and heat it highly, and allow 
a little of the fumes to pass into the experimental receiver, steam will now give 
a very dense fog indeed. 

The effect of dust in producing the cloudy form of condensation of other 
vapours than water was tried. With all the vapours experimented on, which 
included alcohol, benzol, and paraffin oil, it was found that pure air gave no 
clouding whatever, while unfiltered air gave more or less cloudiness with all 
of them. 

The cause of the blue colour of the sky lias long afforded interesting 
matter for speculation. The theory which seems most satisfactorily to explain 
its blue colour depends upon the property which very small particles of matter 
have of scattering only the rays of the blue end of the spectrum, and the 
question is, What are these very small particles composed of 1 It has been 
suggested that they are very small particles of condensed water vapour. Now, 
we have shown the high improbability of water vapour ever condensing out 



35*2 JOHN AITKEN ON 

iii a visible form in pure air, and that if it did condense in those circum- 
stances, the particles would be large. From the all-pervading presence of the 
infinitesimal atmospheric dust, the idea naturally suggests itself, that the blue 
sky may be caused by the light reflected by this dust. What seems to support this 
theory is that, as we ascend to high elevations, the sky becomes deeper blue, 
this being caused by fewer and only the finer of the dust particles being able 
to keep floating in the thin air at these elevations. Further, after rain the 
sky is darker blue, this deepening of the colour being caused by much of the 
dust being washed out by the falling rain. 

I wish now to apply the result of these experiments to the great fog 
question, which Dr Alfred Carpenter opened at the last Social Science 
meeting, and to which at present so much attention is being directed. The 
increased frequency and density of our town fogs are now becoming so great as 
to call for immediate action. But before doing anything, a much clearer 
knowledge of the conditions which produce a fog is necessary, or much time 
will be lost and expense uselessly incurred. I wish, therefore, to call attention 
to the teaching of the experiments described, so far as they bear on this 
important question. What I have to say on this point must, however, be 
received with reservation. The conditions of a laboratory experiment are so 
different and on so small a scale, that it is not safe to carry their teaching to 
the utmost limit, and apply them to the processes which go on in nature. 
We may, however, look to these experiments for facts from which to reason, 
and for processes which will enable us to understand the grander workings of 
nature. 

We have seen that fogs and clouds are produced by the condensation of 
vapour on the dust particles floating in the air. The condensation is 
produced by cold, the result of radiation or expansion of the air, either by 
reduction of barometric pressure or by the elevation of the air into higher regions. 
A fog, therefore, before it appears, is every particle of it represented by a 
particle of very fine invisible dust ; the thick visible fog was previously repre- 
sented by an invisible dust cloud. Now, it is very evident that if there is 
an enormous number of these dust particles in the air, so that they are very 
close to each other, then each particle will only get a very small amount of 
vapour condensed upon it. It will therefore become but little heavier, and 
will float easily in the air. To this light and dense form of condensation we 
give the name of fog. If there are fewer dust particles, then each particle 
gets more vapour, and each particle is heavier and settles sooner. It must 
not be supposed, from this, that rain only falls when these dust particles are 
few, and the vapour particles very large, because there seems to be always 
enough dust in the air to make the cloud particles small enough to keep 
suspended. Their union and fall as rain is determined by certain conditions 



DUST, FOGS, AND CLOUDS. 353 

on which the present inquiry throws no light. But of clouds there are vast 
degrees of texture, the fog being the finest grained, most dense and persistent* 
almost never settling down. 

From this view it will be seen that the vapour condenses on the solid matter 
floating in the air, whether that matter be fine dust or condensed smoke. 
This view I am aware is different from the one generally received, namely, 
that cloud particles are hollow vesicles, hollow to enable them to float, and 
that smoke, &c, attaches itself to the outside of these vesicles. 

Since, then, fogs are produced by an over-abundance of fine atmospheric 
dust in a moist atmosphere, and as we have but little control over the moisture 
in the air, our attention must be directed principally to the diminution of the 
atmospheric dust, if we wish to reduce the density of fogs. We have seen 
that all forms of combustion, however perfect, are great producers of this less 
than microscopic dust. The brilliant flame, the transparent flame, and the 
smoky flame are all alike fog-producers. Perhaps there may be some form of 
combustion which is not a dust-producer, or some form of combustion which may 
give a coarse-grained dust. If there is, it ought to be more generally known. 
As a correction of the present form of combustion, perhaps something could 
be done to arrest the dust before it escapes into the atmosphere. But any 
plan which at present suggests itself is too troublesome and expensive ever to 
be put into general use. To prevent mistakes I may here remark, that when 
speaking of the dust produced by combustion, I do not mean the dust usually 
spoken of in connection with fires, as it is comparatively heavy, and soon 
settles to the ground, nor do I refer to smoke or soot. The dust I refer to is 
the invisible dust, so fine that it scarcely settles out of the air. If we put air 
into the experimental receiver and leave it for days without any communica- 
tion with the outer air, we will still find it fog-producing, though in a very 
marked degree less than at first. 

All our present forms of combustion not only increase the number and 
density of our town fogs, but add to them evils unknown in the fogs which 
veil our hills and overhang our rivers. In the country the fogs are white and 
pure, while in towns they are loaded with smoke and other products of 
imperfect combustion, making the air unwholesome to breathe and filthy to live 
in. But why should these two miseries always come together 1 Either the fog 
or the smoke is bad enough alone ; why should the smoke which usually rises 
and is carried away by the winds fall to the ground when we have fogs % I 
think that the conditions which account for the fog also account for the smoke 
falling. When we have fogs, the atmosphere is nearly saturated with vapour, 
and the smoke particles, being good radiators, are soon cooled, and form 
nuclei on which the vapour condenses. The smoke particles thus become 
loaded with moisture, which prevents them rising, and by sinking into our 

VOL. XXX. PART I. 3 H 



"354 JOHN AITKEN ON 

streets add their murky thickness to the foggy air. This seems to explain the 
well-known sign of falling smoke being an indication of coming rain. That the 
colour or blackness of what is called a pea-soup fog is due to smoke, is, I 
think, evident from the fact that a town fog enters our houses and carries its 
murky thickness into our rooms, and will not be induced to make itself 
invisible however warmly we treat it. It will on no account dissolve into 
thin air, however warm our rooms, for the simple reason that heat only 
dissolves the moisture and leaves the smoke, which constitutes a room fog, 
to settle slowly, and soil and destroy the furniture. If the fog was pure, that 
is to say, was a true fog and nothing but a fog, such as one sees in the 
country, it would dissolve when heated, as every well-conditioned country 
fog does — at least I never remember meeting a fog in a country house. 

But while admitting the bad effects of a fog aggravated by smoke, yet we 
must not forget the probable good effects of the smoke. It has been else- 
where pointed out that the suspended smoke or soot may exercise the well- 
known disinfectant properties possessed by the different forms of carbon. 
Before utterly condemning smoke it will be necessary fully to consider its 
value as a deodoriser. And further, we must remind those who are crying 
for more perfect combustion in our furnaces and grates, that combustion, 
however perfect, will not remove or diminish fogs. It will, however, make 
them cleaner, take away their pea-soupy character, but will not make them 
less frequent, less sulphurous, less persistent, or less dense. 

We have shown that sulphur in its different forms when burned is most 
active as a fog-producer. Now, almost all our coals contain sulphur, which is 
burned along with the coal, and it is certainly worth considering whether some 
restriction ought not to be put on the amount of sulphur in the coal used in 
towns. The quantity of burned sulphur that escapes from our chimneys is 
very great. Suppose we put the amount of coal annually consumed in the 
London district at a little over 7,400,000 tons. Now, the average amount of 
sulphur in English coal is more than 1*2 per cent. Suppose that it is 1 per 
cent., so as to be within the mark, that would give 74,000 tons of sulphur 
burned every year in London fires, or at the rate of about 200 tons in an 
average day, and the amount will be greater in a winter day — a quantity some- 
what alarming, and quite sufficient to account for the density of our fogs. Its 
presence and effects during our fogs is very evident in the discoloured metal 
on our street door and in our houses. 

But, like smoky fires, burnt sulphur is not an unmitigated evil. During 
fogs the air is still and stagnant ; there is no current to clear away the foul 
smells and deadly germs that float in the air, and which might possibly be more 
deadly than they are if it were not for the powerful antiseptic properties of the 
sulphurous acid formed by the burning sulphur. Before condemning the 



DUST, FOGS, AND CLOUDS. 355 

smoke and fog-producing sulphur, it would be well for us thoroughly to in- 
vestigate their saving properties and weigh their advantages, lest we substitute 
a great and hidden danger for an evident but less evil. 

While we look upon fires and all forms of combustion as fog-producers, yet 
we must remember there is ever present plenty of dust in the air to form clouds 
and even fogs ; fires simply increase the amount of the dust. Now it is 
evident that as the rain is constantly washing the dust out of the air, fresh 
supplies must therefore be constantly added. 

We have every reason for supposing that there are immense quantities of 
very fine salt-dust ever floating in the air. This is evidenced by the ever- 
present sodium lime that at one time so troubled spectroscopists. One source 
of the supply of this salt-dust is evidently the ocean, and it affords us another 
example of how very closely the phenomena of nature are interlinked. The 
ocean, which under a tropical sun quietly yields up its waters to be carried 
away by the passing air, almost looks as if he repented the gift, when tossed 
and angry under tempestuous winds, as he sends forth his spray, which dried 
and disguised as fine dust becomes his messenger to cause the waters to cease 
from their vaporous wanderings, descend in fertilising showers, and again 
return to their liquid home. 

Part II. 

Since making my first communication to this Society on Dust, Fogs, and 
Clouds, many of the experiments have been repeated under different conditions 
and with improved arrangements of apparatus. I shall first give a short de- 
scription of the changes made in this direction, which seem to fill up some points 
wanting in the first paper, and shall then describe some experiments made in a 
department of the subject which I have only touched upon. 

We have seen that when steam is blown into dustless air there is no cloudy 
condensation, and that the vapour remained supersaturated till it came in 
contact with the sides of the receiver, on which it deposited itself. My next 
experiments were to determine to what extent dustless air can be super- 
saturated without the vapour condensing into drops — to determine whether 
vapour molecules can combine with one another to form a liquid, or whether 
they must have a nucleus to condense upon even when the vapour is very 
highly supersaturated. It is evidently very difficult to get a definite answer to 
this question, and I shall only describe the direction in which I sought to get 
an answer, the experiments not being sufficiently conclusive to settle the point. 

The first thing to be done was evidently to get quit of all " free surfaces " 
of all nuclei of condensation, and the experiments have resolved themselves 
very much into questions of filtration, as I have not yet arranged any experi- 



356 JOHN AITKEN ON 

merit in which I have been certain there might not have been some nuclei 
present. The first step in this direction was to test the action of the filter 
through which the air passed. All the cotton-wool was removed from the 
filter and a fresh quantity put in. At first only a thin layer was used, and 
its effect tested, noting the degree of cloudy condensation produced. More 
cotton-wool was then put over the first layer, and the improvement noted. 
Fresh quantities were added till no improvemeut was observed. Then double 
the total quantity was put in, and the filter was now considered to be doing all 
that cotton-wool could do to purify the air of the receiver from dust. 

The result was — when a small quantity of steam was blown into the 
receiver there was no cloudy condensation whatever ; the receiver remained 
perfectly clear. But when the steam valve was opened wider and more steam 
allowed to enter, although no effect was noticed at first, yet after a time the 
vapour became so supersaturated that it condensed and fell as fine rain. If a 
still greater amount of steam was blown in, then it was seen condensing on 
entering the receiver, and the falling rainy condensation was seen tossed about 
by the rush of the entering steam. 

Attention was now directed to the steam. It seemed possible that nuclei 
might be given off from the hot sides of the boiler, and from the hot parts from 
which the steam was rising. To prevent any nuclei which might be formed in 
this way from entering the receiver, the end of the steam pipe inside the 
receiver was covered with a cotton-wool filter. The result was, however, as 
before, with little steam, no condensation, with much steam, rainy condensa- 
tion. On account of the tendency of the cotton-wool to get wetted by the 
steam, the action of the filter did not seem satisfactory, some parts getting wet 
and stopping the passage of the steam, and throwing all the duty on the weak 
parts. The experiment was accordingly arranged in the following way : — The 
steam was generated in a glass flask. This flask, filled with water, was placed 
in a vessel full of water, kept boiling during the experiment. In order to make 
the water in the glass flask boil, or rather evaporate, under these conditions, a 
stream of filtered air was blown through it, and the mixture of air and vapour 
blown into the receiver. Again the result was as before — rainy condensation 
when highly supersaturated. By this last arrangement it seems impossible any 
nuclei could be given off from the vessel in which the water was boiled, and 
the fine drops given off by the bubbling of the air and the vapour in the flask 
are probably all caught on the sides of the pipes, because if they did enter they 
would form nuclei in very slightly supersaturated, as well as in highly super- 
saturated vapour. We may therefore conclude from these experiments that 
the nuclei of the rainy condensation in highly supersaturated vapour are either 
some fine form of dust which the cotton-wool cannot keep back, or are pro- 
duced by the vapour molecules combining together without a nucleus. 



DUST, FOGS, AND CLOUDS. 357 

If all nuclei are absent, water may be cooled below the " freezing-point " or 
heated above the "boiling-point" without any change taking place; but there 
seems to be a limit to the amount it may be cooled or heated under these con- 
ditions without the water freezing or boiling. However carefully we may make 
the experiments after the water has been cooled to a certain amount, it always 
freezes without the presence of a free surface, and it also boils without the 
presence of a free surface when heated much above its "boiling-point." In 
these cases there always, however, appears to be some want of continuity or 
uniformity produced by the presence of some substance which exercises an in- 
fluence on the water, and determines a weak point at which the change begins, 
and when once begun progress is of course rapid. In water we can easily under- 
stand how the sides of the vessel and the surfaces of foreign matter, &c, will 
form weak points, from which " free surfaces " are developed, extending into 
the mass of the liquid ; but it is much more difficult to understand how weak 
points can be formed in gases, and even when started they have no power of 
propagating themselves. These considerations would seem to suggest that the 
rainy condensation in filtered air may be produced by some form of nuclei 
which passes the cotton-wool filter, and which are perhaps very small, and do 
not become active as nuclei till a considerable degree of supersaturation is 
attained. 

There are, however, certain considerations which show that if the degree of 
supersaturation is sufficiently great, then condensation will probably take place 
without nuclei. Professor James Thomson * has shown that the isothermal 
curves obtained by Dr Andrews from his experiments on carbonic acid at 
temperatures below the critical temperature of that substance may not be 
really so discontinuous as they appear, and that there may be a condition of 
that substance which would be represented by a continuation of the vapour 
part of the curve beyond the "boiling" or "condensing line." To test this 
point Professor Thomson suggested an experiment in which saturated steam, 
surrounded by a heated vessel, was to be expanded till it was cooled below 
its condensing point for its pressure, and the effect on the volume and pressure 
noted. This experiment, I believe, has never been made. We, however, see 
from the experiments described, that the theoretical extension of the curve 
discovered by Professor Thomson has a real existence. This curve of Professor 
Thomson's shows that the degree of supersaturation possible has a perfectly 
definite limit, beyond which supersaturation is impossible. Further, if we 
examine these curves of Dr Andrews, which we may extend to water, they 
show us that it is only for temperatures below the critical temperature of the 
substance that supersaturation is possible. At temperatures above the critical 

* "Proceedings of the Eoyal Society," No. 130, 1871. 



358 JOHN- AITKEN ON 

temperature there is no boiling and condensation, the change being perfectly 
continuous from the one state to the other, if under those conditions we can 
say there are two states. 

All the previously described experiments have been made at temperatures 
at which the condensed water was in a liquid state. It was now desirable that 
they should be made at lower temperatures, to see if the same conditions are 
necessary when the vapour condenses at temperatures below the "freezing- 
point," and passes from the gaseous to the solid state. The experiments were 
made with the air-pump arrangement of apparatus, the condensation being 
effected by the cooling produced by expansion in the receiver. In the first 
experiments the receiver was placed in a freezing mixture. They were, 
however, repeated under more favourable conditions during the severe cold of 
January last. The apparatus was removed to the open air and experiments 
made with it. The temperature at the time was 8° Fahr. The results were 
the same as at higher temperatures — cloudy condensation with unfiltered air, 
and no condensation when filtered air was used. The amount of cloudiness 
produced was not so great as at higher temperatures. This is due to the 
smaller amount of vapour in the air at the lower temperature. 

I did not succeed in observing any of the optical phenomena produced by 
small crystals of ice in our atmosphere. This was probably due to the 
conditions under which the crystals in the experiment were produced. As the 
crystals were rapidly formed, there would not be time for the vapour molecules 
to arrange themselves in the simpler forms of crystallisation, but by being 
forcibly compelled to solidify, would form complicated shapes, which do not 
give rise to any peculiar optical phenomena. 

In the first part of this paper I have referred to the detection of small 
quantities of matter driven off by heat from pieces of iron, brass, and other 
kinds of matter. By the arrangement of apparatus then described, it was 
shown to be possible to detect the dust drawn off so small a piece of iron wire 
as the x^y of a grain. In later experiments in this direction, the apparatus 
has been entirely changed. In place of using the supersaturation produced by 
mixing steam and cold air, the air-pump arrangement of apparatus has been 
employed, and is found to work much more satisfactorily than the other. The 
impurities drawn off so small a piece of iron wire as the ^nn> of a grain can 
with ease be detected with it. 

The arrangement of the apparatus for this purpose is as follows : — A glass 
flask provided with a tight-fitting stopper, through which pass two tubes, 
which rise to a short distance into the interior of the flask. One tube is 
connected to an air-pump, the other terminates in a stop-cock, to which is 
attached a cotton-wool filter. A piece of glass tube is introduced about the 
middle of the length of this pipe. Some water being placed in the flask, the 






DUST, FOGS, AND CLOUDS. 359 

apparatus is complete. The glass tube must now be thoroughly cleansed. 
This is clone by highly heating it in a Bunsen flame, while air is being drawn 
through it. The end of the glass tube next the filter is now opened, and three or 
four small pieces of iron wire introduced into it. The pieces of wire are placed 
some distance from each other, and near one end of the tube. The tube is 
now closed, and the Bunsen flame placed under the other end of the tube, and 
far enough away from the pieces of iron so as not to heat them. The air in 
the apparatus is now thoroughly cleansed by pumping out the air and admitting 
filtered air, till no cloudiness appears. During this process the height of the 
flame has been reduced, so as the temperature may not be high enough to drive 
anything off the glass tube. When the air is quite pure, and all rainy 
condensation ceased, the flame is reduced to about one-half, so as to leave a 
good margin of safety. After this is done, one of the small pieces of iron wire 
is drawn from the cold part of the tube by means of a magnet, and dropped 
in the hot part, and two or three strokes of the pump are made, to cause a 
current of air to pass through the tube and bring whatever impurities are 
driven off the iron into the flask. The stop-cock at the filter is now closed, 
and a slight vacuum made. The amount of nuclei given off by the wire is 
indicated by the amount of cloudy condensation which now takes place. 

To make further certain that the impurities came from the wire, the piece of 
iron is now removed by means of the magnet, when the filtered air is now found 
to come into the flask without any nuclei, the air remaining cloudless on 
expansion. To make still further certain of the result, another of the pieces of 
wire is drawn into the hot part of the tube, when the cloudiness again appears, 
and again disappears after its removal, or after it has been highly heated. 

The pieces of iron wire experimented on weighed from x^^ to 2 1 o of a 
grain. With pieces so small as this, so abundant and evident is the cloudiness 
produced, that I feel certain that if I could have manipulated, say the xW;^w 
of a grain, the effect would have been perfectly definite and decided. 
Thousands of particles driven off the ^nre °f a grain, and the wire not 
perceptibly lighter afterwards, indicates almost molecular dimensions. It 
seems probable that some of the nuclei in these experiments are driven off as 
gases or vapours. These gases and vapours will afterwards condense when 
cooled in the receiver. It is not necessary that these gases should have nuclei 
on which to condense, as they will be highly supersaturated when cooled to the 
temperature of the receiver, and we know that it is only when supersaturation 
is slight that nuclei are necessary. These gases will, according to their com- 
position, condense either to solid or liquid nuclei, on which the water vapour 
will condense. 

In the first part of this paper attention has been|called to the importance of 
the composition of the atmospheric dust. It was pointed out that some kinds 



360 JOHN AITKEN ON 

of dust will have a greater attraction for water vapour than others, and that 
chloride of sodium dust would probably condense vapour and cause fogging in 
an atmosphere which was not saturated. 

There are evidently two ways in which dust may exert an attraction for 
water vapour, and determine its condensation while still unsaturated. The 
first is the attraction which the surface of some kinds of matter has for vapour, 
a power which they have of condensing a film of water on their surface. This 
power they possess at all degrees of saturation, but the amount they condense 
depends on the degree of saturation. Glass might be taken as an example of 
a substance whose surface has a strong affinity for water, a fact which dis- 
agreeably demonstrates itself in the conducting power of glass insulators of 
electrical apparatus in damp weather. The dust nuclei are so small that the 
condensing power of fine pores is not likely to have any influence. The other 
form of attraction which may exist between the dust and water vapour, is the 
chemical affinity which exists between the two. This will evidently depend on 
the composition of the dust or nuclei. As an example of this form of attraction, 
it will be sufficient here to mention the well-known affinity which chloride of 
sodium and other salts have for water, causing them to become wet when the 
air is moist. 

We shall presently see that besides these two ways in which nuclei may 
condense vapour in unsaturated air, there is another way in which the conden- 
sation may be produced in unsaturated as well as in saturated air without 
nuclei. This happens when there are gases or vapours present which have 
an affinity for each other, and the resulting compound is in a highly super- 
saturated condition. These new compounds under these conditions condense 
and form nuclei, which may be solid or liquid, and may or may not have 
affinity for water. 

Now it is evident that if there are any kinds of matter in the form of dust 
in the air which have an affinity for water vapour, they will determine 
condensation in unsaturated air. Some experiments were made to see to 
what extent cloudy condensation could be produced under these conditions. 
My first experiments were made by burning sulphur, and vapourising chloride 
of sodium. A small quantity of sulphur was lighted, and an open-mouthed 
receiver held over it for a few seconds, and then placed on the table. At first 
scarcely anything was visible, but after a time a decided haze made its 
appearance, and the density of this haze or fog was always in proportion to 
the moisture present in the air. The damper the air the thicker the fogging, 
and if the air was nearly saturated, the result was very remarkable. If the 
inside of the receiver was wetted so as to moisten the air, the sulphur products 
were a little more evident, and on placing the receiver on the table, a thin haze 
could be seen. After a time, however, this haze grew denser and denser, and 



DUST, FOGS, AND CLOUDS. 361 

after fifteen or twenty minutes the receiver was full of a dense white fog, 
which remained for a long time. 

Similar results were got by vapourising chloride of sodium. The salt was 
in some cases vapourised by a Bunsen flame. It was also vapourised by 
placing it on a piece of hot iron, and the receiver held over it to collect the 
vapour, which condensed and formed nuclei, which determined the condensa- 
tion of the water in unsaturated air. In some experiments the salt was 
vaporised in a heated platinum tube and drawn along with air through a coil 
of pipe to cool it, before admitting it into the receiver. In these experiments 
the density of the fogging was in proportion to the vapour present, and if the 
experiment was made in a wetted receiver, the fog took some time to attain its 
maximum density. 

The condensing power of sulphur products and salt can be illustrated in 
another way. The air with either of these substances in suspension, is drawn 
through a coil of pipe to cool it. If now this stream of air is made to strike 
any wetted surface, the wetted surface looks as if it had suddenly become 
heated — a stream of condensed vapour flows away from it. This vapour is, 
of course, invisible if ordinary air is used, and without the powerfully con- 
densing nuclei. 

Experiments on a larger scale were also made with these two substances. 
A little sulphur was burned in a cellar, the air of which was damp, but not 
saturated. The temperature was about 43° Fahr., and the wet and dry bulb 
thermometers showed a difference of from -|° to 1° during the experiments. 
After the sulphur was burned a fogginess was evident, but, on returning half an 
hour afterwards, the fogging was found to have increased very greatly in density, 
the air was very thick, and not the slightest smell of sulphurous acid perceptible. 
This fog hung about the cellar for many hours. The experiment was repeated 
with chloride of sodium, the salt being sprinkled over an alcohol flame. The 
result was similar to the sulphur products, a fogging which gradually increased 
in density, and very slowly cleared away. 

Experiments have also been made by burning sulphur in the open air. 
When the air is dry the fumes can only be traced a short distance, but as the 
amount of moisture increases the cloudiness becomes more and more evident, 
and in certain conditions of the atmosphere the cloudiness can be distinctly 
seen flowing away in the passing air, leaving the sulphur in a pale thin stream 
of vapour, which gradually increases in size and density, and rolls away in a 
horizontal cloudy column, ten or fifteen feet in diameter, clearly marked out 
from the surrounding air. 

There may be a certain amount of doubt as to the action of the heated salt 
in these experiments. When heated in the Bunsen flame it is probable decom- 
position of some of the salt takes place, and part of the result may be due to 
vol. xxx. part i. 3 I 



362 . JOHN AITKEN ON 

the hydrochloric acid set free. In order to prevent this decomposition as much 
as possible, I have made some experiments at as low temperatures as possible, 
and the results are the same as when higher temperatures are used, allowance 
being made for the smaller amount of salt volatilised. 

The action of the products of combustion of sulphur would appear to be 
something like the following : — When the sulphur combines with the oxygen of 
the air, sulphurous acid is formed. I have shown in the first part of this paper 
that sulphurous acid has but little condensing power ; we must therefore look 
to the change which takes place in the sulphurous acid for the explanation of 
the wonderful condensing power of the sulphur products. The sulphurous acid 
becomes further oxidised in the air, and sulphuric acid is produced, and it is 
the great affinity which this sulphuric acid has for water which enables it to 
rob the air of its moisture and condense it in visible form. It does not seem 
to take long for the sulphurous to change to sulphuric acid in the air. A 
short time after the sulphur was burned in the cellar all smell of sulphurous 
acid was gone, and I am informed by Dr Wallace that he has found that all 
traces of sulphurous acid cease at a short distance from calcining ironstone 
bings in which much sulphur is being burned. The gradual thickening of the 
sulphur fog will probably be in part due to this gradual change of sulphurous to 
sulphuric acid. The gradual thickening of these fogs is also in part due to 
the slow evaporation of the water from the sides of the receiver, and subsequent 
condensation on the absorbing nuclei. 

I find that the fumes from highly concentrated sulphuric acid have a fog- 
producing power similar to the products of combustion of sulphur. If we 
highly heat a glass rod wetted with sulphuric acid, or heat the acid in a 
platinum cup, and admit a little of the fumes into the receiver, they are found 
to have a very strong fog-producing power. 

The above represents something like what the action of sulphuric acid is in 
moist air, in which there are no other vapours or gases with which this acid tends 
to combine. Before considering these more complicated effects I shall describe 
some experiments made to test the action of acid vapours on moist filtered air. 
The apparatus consisted of the air-pump arrangement, with test receiver or 
flask, one pipe as before being connected with the air-pump, and the other with 
the filter. Between the receiver and the filter was placed a test tube, in which 
was placed the acid to be experimented upon. The filtered air was caused to 
bubble through the acid on its way to the moist air in the receiver, the acid 
being generally kept at the temperature of the room. 

When nitric acid is put in the test tube and filtered air passed through it, it 
is found that its vapour always gives rise to fumes when mixed with the moist 
air in the receiver. These fumes — as cloudy condensation in unsaturated air 
may be called — may therefore be produced without nuclei when nitric acid is 



DUST, FOGS, AND CLOUDS. 363 

used. When the air in the receiver is expanded and cooled, this cloudy con- 
densation becomes thicker. 

When commercial hydrochloric acid is put in the test tube, its vapour does 
not give rise to fumes on mixing with the moist air in the receiver, and on ex- 
panding and cooling the air, no fumes appear, only the rainy form of conden- 
sation is produced. A quantity of very strong hydrochloric acid was prepared 
by keeping the solution in which the acid was condensed in a freezing mixture. 
This acid fumed abundantly in the air, but gave no fumes in filtered air, and 
only rainy condensation when the pressure was reduced. 

These two acids act very differently, the first condensing freely at many 
centres, and without nuclei, and giving a foggy condensation in pure and unsatu- 
rated air, while the hydrochloric acid only condenses with difficulty, and at few 
centres, and only gives the rainy form of condensation when supersaturated. 

The next experiments were made with commercial sulphuric acid, and also 
with some of the acid concentrated by boiling in a glass vessel. The. air which 
had passed through this acid gave no fumes, but on making the slightest ex- 
pansion a fog appeared. This fog is quite characteristic of sulphuric acid, and 
is quite different from any artificial fog I have seen. The particles are 
extremely small, and the display of colour remarkably brilliant, and when 
properly lighted rivalling in distinctness the colours of the soap bubble. This 
beautiful fog is only got when the acid is strong, and I think is best produced 
when the entering air is dry. This point, however, requires confirmation, though 
the result might be expected, as the surface of the acid will then be less 
weakened by moisture abstracted from the air. After the acid has absorbed 
much vapour, or if water has been added to it, the fogging decreases and gives 
place to the rainy form of condensation when expansion is made. This rainy 
condensation also disappears when the acid is very weak. If we heat the strong 
acid to a temperature of about 60° or 70° C, the vapour condenses and forms 
fumes in pure air without nuclei, and without being expanded. 

These experiments show that water vapour may be condensed without 
nuclei being present. The affinities which the vapours of the acids have for the 
water, causing the formation of new compounds, and these compounds being 
highly supersaturated, condense easily without nuclei, and in certain circum- 
stances this condensation may be determined in even unsaturated air. These 
water-acid nuclei once formed, continue to act as centres of condensation. In 
these cases the manufactured nuclei are liquid, but solid nuclei may be formed 
in a similar manner. This may be shown by the following experiment. Place 
hydrochloric acid in the receiver or flask, and pump out all the air and replace 
it with filtered air. If, after this is done, and the acid shows no sign of cloudi- 
ness, and nothing but rainy condensation on expansion, we take the stopper 
out of a bottle of ammonia and hold it near the filter, so that the escaping 



304 JOHN AITKEN ON 

gaseous ammonia may pass along with the air through the filter, the ammonia 
on arriving in the flask will combine with the hydrochloric acid and form a 
dense cloud of sal-ammoniac. When the ammonia and the hydrochloric acid 
combine in the filtered air, the tension of the sal-ammoniac vapour so formed 
is enormously greater than that due to the temperature, and it easily condenses 
without nuclei. This experiment suggests that part of the rainy condensation 
given by hydrochloric acid may be due to the ammonia in the air combining 
with the acid and forming sal-ammoniac nuclei on which the vapour condenses. 

These experiments show how nuclei may be formed from gases in the air, 
and these nuclei may have so great an affinity for water vapour as to cause 
it to condense on them from an unsaturated atmosphere. 

Returning again to the action of the products of combustion of sulphur in 
air, we have seen that these products alone can determine the condensation of 
water vapour from unsaturated air. There are, however, many substances in 
the air with which this acid will tend to combine. It would be impossible 
to go over all the substances in the air which have affinities for this acid, 
and consider the effects of these new compounds, in moist air. I have, how- 
ever, selected one, which from the magnitude of its effects deserves special 
notice. That substance is ammonia, another of the products of combustion of 
our coal fires. If we take an open-mouthed receiver wetted on the inside, 
and hold it over a little burning sulphur for a few seconds, as in the previous 
experiment, we will get a thin haze, which we know tends to thicken. But if 
on placing the receiver on its tray, we put a drop of ammonia on a piece of 
glass and introduce it into the receiver, the result is very striking. Dense 
fumes will be seen to rise from the ammonia, and in a few minutes the receiver 
will be full of a fog so thick it will be impossible to see an object in the middle 
of the receiver. In this case there are evidently formed solid nuclei, composed 
of sulphite and sulphate of ammonium, in a very fine state of subdivision. 
The intense cloudiness is only in part due to this solid, the greater part is due 
to the condensation of water vapour. If the experiment is made in dry air the 
fogging is not nearly so intense as in moist air. By burning a larger amount 
of sulphur in the moist air of the receiver, we can easily make a fog so very 
intense that it is impossible to see through an inch of it. This fog is found to 
be very suitable for experiments on vortex rings, as it is easily prepared, and 
the " dead " rings dissolve, and do not thicken the air of the room to the same 
extent as the usual sal-ammoniac rings. 

Experiments were also made in the cellar with this fog-producer. The 
wet and dry bulb thermometers at the time showed a difference of fully one 
degree. Yet by burning a few grains of sulphur, and dropping on a piece of 
paper a little ammonia, the cellar became filled with a most intense fog, 
many times more intense than would be produced by the sulphur alone. 



DUST, FOGS, AND CLOUDS. 365 

Using the same apparatus as was used for determining the fuming power 
of the different acids in filtered air, it is found that when experimenting on 
sulphuric acid and vapour of ammonia, that sulphate fumes are formed in the 
receiver if the acid is slightly heated, thus showing that this sulphate dust can 
form without nuclei. It, however, seems in the highest degree probable that 
when dust is present the dust particles will form the centres on which the 
sulphate will condense. 

Almost all salts when heated in a Bunsen flame produce nuclei which 
determine condensation in unsaturated air. The condensing power of the 
different products, however, differ greatly. The bicarbonate of soda gives but 
little effect, while chloride of calcium and bromide of potassium are much 
more powerful. But by far the most powerful artificial fog-producing substance 
when used in this way is the chloride of magnesium. If we put a small 
quantity of this salt on a piece of wire-cloth, and heat it with the Bunsen flame, 
and collect the products in a wetted receiver, the fog will be seen rapidly 
forming and showering down the sides of the receiver. As rapidly as the 
water is evaporated from the sides of the receiver it is condensed by the active 
nuclei in the gases. After the receiver has been placed on the table for a few 
minutes it will be found full of a fog so dense it is only possible to see through 
a depth of five centimeters of it. When a little of this chloride was heated in 
an alcohol flame in the cellar the result was a fog many times more dense than 
that produced by sulphur alone. The fog-producing power of the heated 
chloride of magnesium would appear to be due to the salt being decomposed 
by the heat, and free hydrochloric acid being driven off in a highly concentrated 
state. The amount of hydrochloric acid is, however, small considering the 
density of the resulting fog. The density of this fog is very much greater than 
the fog produced by hydrochloric acid, prepared from chloride of sodium and 
sulphuric acid. 

In all these cases the reactions are excessively difficult to trace. Other 
experiments in which the action is much simpler were made by burning a little 
sodium in the receiver. The combustion of this substance gives rise to its 
oxides in a fine state of division. This fine soda-dust when mixed with dry 
air gives but little cloudiness, but when mixed with damp air a dense fogging 
results. Potassium when burned gives a similar effect, but the fog is not so 
intense. 

We may conclude from these experiments — 1. That as regards cloudy con- 
densation of vapour in our atmosphere there is dust and dust. Some kinds of 
dust have such an affinity for water that they determine the condensation of 
vapour in unsaturated air, while other kinds of dust only form nuclei when the 
air is supersaturated, that is, they only form free surfaces on which the vapour 
may condense and prevent supersaturation. In many of the experiments it 



366 JOHN AITKEN ON 

was noticed that when the air was nearly purified, when all the dust which had 
an affinity for vapour had received its burden of water and settled down, that 
there remained to near the end of the experiment some particles which seemed 
to require a certain degree of supersaturation before they became active. In 
highly supersaturated air all kinds of dust will form nuclei and determine 
condensation, but in unsaturated air only those kinds of dust which have an 
affinity for water will be active. We have precisely corresponding phenonema 
to this in freezing, melting, and boiling. We have water in a solid state at a 
temperature above the "melting-point," when it is combined with some other 
substance, as in the water of crystallisation of salts. Water may be liquid at a 
temperature below the " freezing-point " when mixed with some salts. Water 
boils at a temperature above its " boiling-point " when it holds some salts in 
solution, and boils below its " boiling-point " when mixed with some substance 
having a lower " boiling-point " than water. 

2. This affinity which some kinds of dust have for vapour explains why it 
is that our breath and escaping steam dissolve even in foggy air. The large 
cloudy particles in our breath and in condensed steam tend to evaporate in the 
same air in which condensation is taking place, because the dust particles on 
which the breath has condensed have had their affinities more than satisfied, they 
therefore tend to part with their surplus by evaporisation in the same air as those 
particles which have not had their affinities satisfied tend to condense it. 

3. Dry fogs are produced by the affinity which the dust particles have 
for water vapour, in virtue of which they are enabled to condense vapour in 
unsaturated air. From the experiments with chloride of sodium, from the 
known affinity of that salt for water, and from the fact that great quantities of 
salt-dust are ever present in the air, it is evident that if it is not the cause of 
dry fogs in the country it must play some part in those phenomena. There will 
doubtless be other kinds of nuclei having affinities for water which will cause dry 
fogs. The nature and composition of these other nuclei will probably be best 
arrived at by collecting the fog particles by washing or otherwise, and analysing 
them. 

4. That as the products of combustion of the sulphur in our coals, espe- 
cially when mixed with the other products of combustion, such as ammonia, 
have the power of determining the condensation of water vapour in unsaturated 
air, and give rise to a very fine-textured dry fog, they are probably one of 
the chief causes of our town fogs, as they have a greater condensing power than 
the products of combustion of pure coal. 

Though there may seem to be but little doubt that products of combus- 
tion when mixed with the sulphur compounds are most active producers 
of town fogs, yet we must not rest satisfied that they explain everything. 
There may be other causes at work, and conditions yet requiring explanation, 



DUST, FOGS, AND CLOUDS. 367 

but as these involve intricate chemical reactions, it will be advisable that the 
matter be now handed over to the consideration of the chemist. 

These chemical nuclei, as they might be called, though found in far greatest 
abundance in the air of our towns, will no doubt be also found in the air of 
the country. We know that sulphuric acid and ammonia are constantly being 
produced by decomposing animal and vegetable matter, and we know that 
these substances, along with nitric acid and other gases and vapours, are 
always present in the air. 

Again, we have the gases given off from volcanoes, and the amount from 
this source must be considerable. There are about two hundred active vol- 
canoes constantly discharging their gases into our atmosphere, and it has been 
roughly calculated that volcanoes evolve ten times more carbonic acid than is 
given off by the combustion of all kinds of carbonised material. With this 
carbonic acid there is given off great quantities of sulphurous and other gases 
which will condense and form nuclei. 

Vegetation, both when alive and when dead, gives off vast quantities of 
small organic particles, and microscopic life, which almost seem to populate 
the air we breath, and will of course add much to the dust in our atmosphere. 

Professor Tyndall has shown that light decomposes certain gases and 
vapours, and that this decomposition is greatly aided by the presence of 
other gases or vapours. It seems therefore probable that the sun's rays will 
decompose some of the gases and vapours in the air, and if these decomposed 
substances have a lower vapour tension than the substance from which they 
are formed, they condense into very fine particles. These particles may be 
solid or liquid, and will form nuclei for the condensation of water vapour. 

We know that there are ever present in our atmosphere great quantities of 
chloride of sodium and other kinds of dust which have affinities for water. 
These dust particles by their affinities for water vapour cause condensation 
to take place in unsaturated air, and if present in great quantities give rise 
to dry fogs. ' Let us look briefly at the effect of this affinity between the dust 
and the vapour. If there was no affinity between the two, then condensation 
would only begin when supersaturation began, and those dust particles which 
permitted the vapour to condense on them easiest would get most vapour, and 
would tend to grow largest. This would evidently tend to inequality in the 
size of the cloud particles which would determine the fall of some of them 
through the others. But if there is an affinity between the dust and the vapour, 
then each particle of dust tends to take the same amount of vapour, and if one 
particle gets more than its proportion, the others tend to rob it of its surplus. 
This evidently tends to equality in the size of the cloud particles, and tends 
also to prevent any of them falling through the others, and thus prevents it 
beginning to rain, that is, if rain drops are formed by the collision and union of 



368 JOHN AITKEN ON DUST, FOGS, AND CLOUDS. 

the quickly foiling particles with those falling more slowly. It would thus 
seem that while on one side if we have no dust we would have no clouds and 
probably no rain, as we don't know whether the air would ever become suffi- 
ciently supersaturated to condense without nuclei. On the other hand, an 
over-abundance of dust having affinities for water vapour also prevents the 
vapour falling as rain, as the vapour under these conditions condenses into 
minute particles which all tend to be of equal size, and none of them are able 
to fall quickly enough amongst the others to cause collisions. The result is the 
condensed vapour cloud instead of falling in minute parts as rain, tends to fall 
as a whole. The air becomes so loaded with the water held in mechanical 
suspension that it is dragged downwards by its weight. If we make artificial 
fogs with sulphur fumes and ammonia, or by heating chloride of magnesium, 
the fog is so heavy it can be poured from one vessel to another. 

After the affinities of the dust particles are satisfied, this tendency to 
stability no longer exists. After this stage the growth of the particles becomes 
unequal, and, as has been shown by Professor Clerk Maxwell,* the larger 
drops or particles in a cloud tend to rob the smaller ones, or rather, from what 
we now know, will tend to prevent them growing after the affinities of the 
nuclei are satisfied. 

It would appear, then, that condensation will always begin in our atmosphere 
before the air is saturated. There is, however, still much to be done in this 
department of our subject to determine whether the amount of cloudy conden- 
sation is always the same for the same degree of saturation, or if it varies ; and 
if it varies, to find the composition and source of the nuclei which cause the 
variations. 

I feel that these two papers only start this inquiry. Much, very much, still 
remains to be done. Like a traveller who has landed in an unknown country, 
I am conscious my faltering steps have extended but little beyond the starting- 
point. All around extends the unknown, and the distance is closed in by many 
an Alpine peak, whose slopes will require more vigorous steps than mine to 
surmount. It is with reluctance I